This answer and discussion in comments below this answer mention that for an ion of mass $m$ and charge $q$ accelerated by a voltage $V$ the momentum it receives (impulse) is

$$p = \sqrt{2mqV} = \sqrt{2mE}$$

and the mass-specific impulse for one atom would be that divided by mass:


This suggests that if you used 4He+ or 1H+ in an ion thruster or engine you could get about 5.7 or even 11.5 times more Isp compared to using 131Xe+ ions.

Xenon and krypton are popular despite their heavy mass because they are simply much easier to

  1. put in bottles
  2. ionize in the kinds of plasma conditions that are convenient to make on a small spacecraft
  3. they are not very reactive with the materials used in the engines.

Has the "ion sorcery" for light gases like hydrogen and helium been explored experimentally for future ion propulsion technology? What about neon at least?

Just fyi iodine has also been explored because while heavy (bad) and easy-ish to ionize (good) like xenon, it can be stored as a solid and sublimated on-demand. While storing large quantities of liquid helium for long flights will be a challenge and require a sun shade, liquid and solid sources of gaseous hydrogen and hydrogen-containing gases are probably within reach.

  • 3
    $\begingroup$ People make design choices for the performance of the mission by the craft. Fuel choices are made to maximize that. Heavier atoms give more thrust per energy (see your first equation) and that’s what the designer cares about when craft mass and power are limiters. Specific impulse matters when fuel mass is large and (combustion) energy comes from the fuel, but that’s not the ion case. $\endgroup$ Commented Apr 13, 2020 at 6:31
  • $\begingroup$ @BobJacobsen The correct answer to "Have light gases like hydrogen or helium been explored for ion propulsion?" is "Yes they have!" and a good answer will explain it. Not every possible future mission will be clone of DAWN with DAWN's constraints. Imposing that seems counterproductive. Answering "No nobody has or would ever explore hydrogen or helium because all missions will be just like DAWN forever" seems to probably not be true. $\endgroup$
    – uhoh
    Commented Apr 15, 2020 at 13:28
  • 1
    $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – called2voyage
    Commented Apr 15, 2020 at 13:47
  • $\begingroup$ But better ISP, this mean also less thrust per energy, isn't it? Although as an independent variable, more ISP is better, on a global mission profile, that choice could mean slower time to destination because thrust is lowered, I think. $\endgroup$ Commented Apr 6, 2021 at 7:47
  • $\begingroup$ Isp reflects thrust per rate of mass use. $\endgroup$
    – uhoh
    Commented Apr 6, 2021 at 8:15

3 Answers 3


(Top edit: The Question asserts "Xenon and krypton are popular despite their heavy mass" and asks about exploring H or He ion propellants for improved Isp. This answer shows that lighter is not better for ion thrusters, because Isp is not the proper measure of a power-limited situation. Hence, although lighter atoms have been explored for other reasons, they're certainly not explored because they provide better Isp.)

Typical ion thrusters have a small mass of propellant compared to the mass of the power generation system plus the rest of the spacecraft. In that case, the goal is to get as much thrust from the ions as possible given the power that's available.

Referring to the first equation from the Question:

$$p = \sqrt{2mqV} = \sqrt{2mE}$$

for a fixed amount of energy E, the greatest outgoing momentum hence greatest thrust comes from a larger mass atom. Switching from H to Xe is about a $\sqrt{131} \approx 12$ times increase in thrust, at the price of adding a couple kilograms to a much more massive spacecraft.

It's true that a heavier atom is ejected slower, as $E =1/2 m v^2$ means $v = \sqrt{2E/m}$. But that's more than made up for but the larger $m$ in $mv$.

Dawn is beyond the small-thruster regime into the ion engine region. It launched with 425kg of Xe on a 750kg spacecraft.

The Dawn spacecraft carried 425 kilograms (937 pounds) of xenon propellant at launch. Xenon was chosen because it is chemically inert, easily stored in a compact form, and the atoms are relatively heavy so they provide a relatively large thrust compared to other candidate propellants.

(Quote on this Dawn page)

The same number of H atoms would be only about $425/130 = 3.3 \rm{kg}$. But with the power available, the thrust would go down by a factor of 12 (although acceleration drops a bit less, as average total mass has gone down by about a sixth). That would have adverse impact on the mission. And the only way to restore the original thrust hence acceleration with H fuel would be to increase the size of the power provided by a similar factor of about 12. Dawn's solar arrays (which power the entire craft, not just the engines) are $18\rm{m}^2$ now; you'd be adding another $100\rm{m}^2$ or more, with consequent increase in mass, need for more thrust, etc. In discussion, it's argued that what matters is the velocity of the exhaust, not the momentum. This is only true in a specific approximation where the energy of the outgoing exhaust is not intrinsically limited by some other process. For example, if you're combusting 10kg of LOX LH2, then you want that mass to be ejected with as much speed as possible using as much of the combustion energy as possible. For a constant mass (flow), it is speed that matters. But ion propulsion is (so far? usually?) limited by available power, which is a different regime. You can't compare two different mass flows without taking into account how much the available power can accelerate them.

So how does the power limit come in? Here, a higher velocity of the charged particles in the exhaust works against you. The current is $qv$, so the power needed is $qvV$: Higher velocity is more energy needed per unit of charge. Since you're limited by the energy you can put in to the exhaust stream, the exhaust velocity is effectively fixed for the thruster.

Analytically, the available power is given by voltage and current (capital letters are electric quantities, lower case are mechanical, the $i$ subscript is per-ion): $$ P = I V$$

Break down current into total charge per second and velocity:

$$ P/V = I = q_i dN_i/dt v$$

where $dN_i/dt$ is the number of ions exhausted per second. Expressing this in terms of the ion's intrinsic charge to mass ratio:

$$ P/V = I = (m_i dN_i/dt) q_i/m_i v$$

where the term in () is the total mass exhausted per second. Regrouping to highlight momentum:

$$ P/V = q_i/m_i (dm/dt) v$$

$$ P/V = q_i/m_i dp/dt $$

$dp/dt$ gives the thrust, so finally:

$$dp/dt = P/V (m_i/q_i) $$

More power and higher mass ions lead to more thrust; more specifically a higher mass/charge ratio is better.

  • 3
    $\begingroup$ This is exactly the correct answer - it just doesn't make sense to use light gases because of power requirements to get the same acceleration. Nobody cares about dv alone, net acceleration is the crucial factor. $\endgroup$
    – asdfex
    Commented Apr 13, 2020 at 9:09
  • 3
    $\begingroup$ @uhoh I don't see a wrong statement there. Heavy atoms are better. There is no way for any reasonably-sized, close-future technology probe to be better off using light ions. $\endgroup$
    – asdfex
    Commented Apr 13, 2020 at 11:12
  • 3
    $\begingroup$ No, that's wrong. Xenon is chosen because it's heavy. $\endgroup$
    – asdfex
    Commented Apr 13, 2020 at 13:15
  • 2
    $\begingroup$ This is the correct answer. Note that the energy $E$ is provided by the ion accelerator, and is the same regardless of ion mass (although increasing ion charge can increase the ion $E$, at the cost of not having nearly as many high charge state ions without a lot of trouble on the source side). Since momentum is key (conservation of momentum in the inertial frame), mass is good. $\endgroup$
    – Jon Custer
    Commented Apr 13, 2020 at 17:50
  • 4
    $\begingroup$ I'm puzzled by the claims that this is the correct answer, given that it doesn't seem to address the title question at all. $\endgroup$ Commented Apr 13, 2020 at 19:08

I will write another complementary answer because the top one unfortunately has some unit and physics inconsistencies.

First to answer the question:

have light gases like hydrogen or helium been explored for ion propulsion?

Depends on what you call "ion propulsion". If you mean gridded ion thrusters, then most probably no (see below). On the other hand, other types of electric propulsion concepts do use hydrogen and helium.

Electrothermal devices, such as resistojets and arcjets, generally have a maximum operating temperature due to material constraints. Their specific impulse can be very roughly estimated as $I_{sp} = \frac{1}{g_0} \sqrt{2 c_p T}$, where $g_0$ is the standard gravity acceleration, $c_p$ is the constant-pressure specific heat capacity and $T$ is the temperature. Since $c_p = C/m$, where $C$ is the heat capacity and $m$ is the propellant molecule mass, to increase the $I_{sp}$ to desirable levels (close or higher than 1000s), it is necessary both to increase $C$ by using atomic particles (decreasing the degrees of freedom of the particle), and decrease $m$.

Electromagnetic devices, such as the MPD were also largely operated with light propellants. MPDs specifically use it to increase the $I_{sp}$ at very high power levels. Lithium is particularly preferred to mitigate the erosion of the cathode which is one of the major problems of this technology.

Now to expand a bit on why it is not desirable to use Hydrogen or Helium (or any other light propellant) with gridded ion thrusters (or Hall thrusters).

Consider first that an ion thruster generates a collimated beam, with just one ion species. The power required to generate this beam is $P_b=I_bV_s$, where $I_b$ is the ion current and $V_s$ is the screen grid voltage.

The kinetic energy at which the ions leave the thruster is approximately $E_i=\frac{1}{2}m_iv_i^2$. Since the ions are accelerated by a electrostatic potential profile, the energy is equivalent to $E = q_iV_s$, since $V_s$ is approximately the potential at which is the ions are created.

The ion beam current can be rewritten as $I_b = \frac{q_i}{m_i}\dot{m}_i$, where $\dot{m}_i$ is the mass flow rate of ions. Using these four equations, the power can be rewritten as

$$ P_b = \left( \frac{q_i}{m_i} \dot{m}_i \right) \left( \frac{1}{2} \frac{m_i}{q_i} v_i^2 \right) = \frac{1}{2} \dot{m}_i v_i^2$$

which is exactly the formula for the "jet" or "beam" power given in textbooks. As you can see there is no guarantee that if you change the mass or charge of the ions the power consumed to accelerate the beam will change. The only thing that matters in this case is the mass flow rate and the speed that the ions exit the system.

On the other hand, we can rewrite the equation to evidence the thrust, as

$$P_b\approx \frac{1}{2} Tv_i \approx \frac{1}{2} T \sqrt{\frac{2q_iV_s}{m_i}}$$

In this case, for a fixed screen grid voltage $V_s$, to increase the thrust one has to either increase the ion mass or decrease its charge. So if you have a system that has a maximum voltage limitation (due to arcing for example) you have to increase the mass of ions to guarantee that you have a reasonable level of thrust thus decreasing your mission time.

Furthermore, as shown in the first chapter of Jahn's book, it is also important to notice that a higher $I_{sp}$ is not always better. There is an optimal specific impulse for a given mission which can be roughly estimated by

$$ \hat{I}_{sp} \approx \frac{1}{g_0} \sqrt{\frac{2 \eta \Delta t}{\alpha}} $$

where $g_0$ is the standard gravity acceleration, $\eta$ is the thruster efficiency, $\Delta t$ is the mission time and $\alpha$ is the ratio of the power processing subsystem and the power required by the thruster. This happens because as you increase the $I_{sp}$ you also increase the size of your power system, decreasing the overall payload mass of your spacecraft. This way, using a very light propellant like hydrogen to give you the maximum possible $I_{sp}$ may in fact just increase the overall required mass of your spacecraft.

The last point to consider is the amount of power you spend to create the ions. Consider that every time you ionize a neutral atom you spend an ionization energy $E_{iz}$. This way, the power you spend to ionize the ion mass flow rate is $P_{iz}= E_{iz}\frac{\dot{m}_i}{m_i}$. Using the expressions for both power terms (and excluding additional power contributions from the system), you have

$$ P = \dot{m}_i \left( \frac{v_i^2}{2} + \frac{E_{iz}}{m_i} \right) $$

The first ionization energy for xenon is 12.1 eV and of hydrogen is 13.6 eV. So you will spend around 150 times more power to ionize the same mass flow rate when using hydrogen as compared to xenon. Of course this is a very simplified analysis and more precise numbers would require a detailed analysis of the reactions and discharges using both propellants.

As commented by @asdfex, it is also important to notice that the ratio of the beam to ionization power will be constant in respect to the ion mass,

$$\frac{P_b}{P_{iz}} = \frac{qV_s}{E_{iz}}$$

This way, if you fix the screen grid voltage $V_s$ and let the average ion speed to vary, the only method to increase the amount of power going to the beam when compared to the ionization is by choosing propellants that have a smaller $E_{iz}$.

  • $\begingroup$ This is really interesting! "Optimal specific impulse" is new to me (as far as I can remember at least). It seems then plausible that there can be some missions such that the higher Isp of a lighter species would be of some advantage; for example a deep space mission that could take a long time and have the right balance of solar to nuclear thermal power. Do you have any thoughts on the question itself? "Have light gases like hydrogen or helium been explored for ion propulsion?" which the "top answer" didn't bother to answer either. $\endgroup$
    – uhoh
    Commented Sep 7, 2022 at 15:27
  • 1
    $\begingroup$ Regarding the last point: The ratio of ionization energy to acceleration energy is constant - for each atom you need ~12 eV for ionization and put ~ 2keV in for acceleration (assuming the same 2kV potential in both thrusters) $\endgroup$
    – asdfex
    Commented Sep 7, 2022 at 17:02
  • 1
    $\begingroup$ @uhoh sorry I haven't payed attention, I'll edit the question to add some comments. $\endgroup$
    – grafo
    Commented Sep 7, 2022 at 19:30
  • $\begingroup$ There are some resources available here, but I'm not sure you are able to access it; space.stackexchange.com/a/51291/12102 $\endgroup$
    – uhoh
    Commented Sep 7, 2022 at 20:21

Have light gases like hydrogen or helium been explored for ion propulsion?

I'll quote @BobJabobsen:

Whatever, dude. The 2nd to last paragraph of your first cited paper makes the point quite well too. Have they been explored? Yes, and even though they have better Isp, they've been rejected because they're worse for an actual vehicle due to power (not just ionization). I'll leave the rest to the up and down votes.

Which is baffling because that's all that that answer ever says. Lighter gases will have better Isp, and so they might have been explored. The the indignation comes from pretending a different question was asked than what was actually asked, and the highly upvoted answer pretends that a different question was asked.

The question asks if they have been explored, and includes an explanation for why one might want to.

Rather than answer the question as asked with facts, there is a lot of chauvinistic opining why one would never do this in answers and in comments, and an inordinate amount of down voting on my answer post that elaborates on the premise.

Basically folks wanted to write answers to "Why didn't NASA use helium for Dawn?" and so they pretended that's what my question asked, which it did not.

There are other kinds of missions, other kinds of applications. Some orbits need only a tiny amount of delta-v for station keeping per year and have plenty of sunshine.

Just for an illustrative example, JWST will only need 2 to 4 meters per second of delta-v per year for station keeping. 1, 2 A cubesat in a similar orbit with some run-of-the-mill cubesat solar panes will have plenty of power. What it needs to be careful about is Isp and mass.

Not every spacecraft is a Dawn clone!

Arguments boil down to spacecraft being energy-limited rather than mass-limited, so if you are in a hurry you want to use heavier species. That's light-years away from what the question asks.

So I say the positions here are chauvinistic opining because they are all based on all space missions being exactly like historical ones; moderate distances and little patience.

Well, have they?

Of course they have!


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.