Like why use RTGs (Radioisotope thermoelectric generator) to power small spacecraft instead of just placing batteries inside them to power stuff?

  • 32
    $\begingroup$ Batteries were and are still used… but not for long-duration missions where you can't use alternative power sources like solar arrays. $\endgroup$ – DarkDust May 15 '17 at 15:03
  • 13
    $\begingroup$ See also List of spacecraft powered by non-rechargeable batteries $\endgroup$ – DarkDust May 15 '17 at 15:04
  • 9
    $\begingroup$ Many spacecraft in low Earth orbit that use solar arrays but regularly have the Sun eclipsed by the Earth use batteries so they can operate during those eclipse periods. On the other hand, recharging is not an option for vehicles that go beyond Jupiter. $\endgroup$ – David Hammen May 15 '17 at 15:51
  • 25
    $\begingroup$ I think it would be helpful to clarify that a RTG is a power source, whereas batteries are for power storage. For satellites, batteries are used to store excess solar-generated electricity for time when solar power is not available. RTGs are power sources used when reliance on solar-generated power is not feasible. $\endgroup$ – JS. May 15 '17 at 17:36
  • 10
    $\begingroup$ Even some (most?) RTG-powered spacecraft also use batteries, because the power consumption has a peak-to-average power ratio significantly greater than 1. The battery supplies peak loads and is gradually recharged by the RTG. $\endgroup$ – pericynthion May 15 '17 at 17:58

Energy density for:

NiMH C battery -> 237,073 Joules per Kilogram.

Plutonium 238 (used in RTGs) -> 2,239,000,000,000 Joules per Kilogram.

Even if we assume that only 10% of a RTG weight is actually Plutonium, then we still get about 9,400,000 times as much energy available as heat from an RTG as from the same mass of batteries.

In most deep-space missions, landers, and rovers, heat generation is essential to maintain spacecraft function. However, for electrical power, the best conversion efficiency from RTG thermal to electrical power is about 7%, making it "only" 661,000 times as much energy available as electrical power from a RTG as from the same mass of batteries. That's still a pretty huge difference!

Source: Energy Density, wikipedia

  • 3
    $\begingroup$ xkcd.com/1162 (talking about uranium, not plutonium) $\endgroup$ – Andrew Grimm May 18 '17 at 0:41
  • $\begingroup$ @AndrewGrimm Likely also talking about a chain reaction rather than radioactive decay. $\endgroup$ – JollyJoker May 18 '17 at 10:48
  • 1
    $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$ – called2voyage May 18 '17 at 13:01
  • 1
    $\begingroup$ @AndrewGrimm xkcd.com/2115 $\endgroup$ – Speedphoenix Jul 14 '19 at 14:00

RTGs are used in a very small number of spacecraft. They are used only when there are no other options, i.e. for long missions too far away from the Sun to make solar panels feasible.
Those missions have requirements for a few hundred W of power continuously over a decade or more. If you were to use chemical batteries to supply that much power, your spacecraft would become too heavy to launch.


Nuclear decay is just simply the most energy dense fuel there is. This is enough to overcome huge inefficiencies in power conversion. We can even ignore the inefficiencies of alternate storage methods, and still conclude that fissile material will store more energy per unit mass.


For RTGs I'll refer to Wikipedia's article about RTGs that have previously been or are in service. By their nature, the output and efficiency of an RTG is complicated to compute, so I'll refer to the actual measured power output, and assume the output declines along with the nuclear decay (half life of 87 years for 238Pu).

Evaluating the RTGs used in aerospace applications, the absolute worst performer was the SNAP-3B generator with a specific power of 1.3 W/kg (at launch). This was used on the Transit4B satellite, which was operational for roughly 1 year (accidentally destroyed by nuclear test detonation). During this time its 2.1kg RTG produced roughly 23.558 kWh of electricity. This gives a specific storage of 11.2 kWh/kg


Typical quoted values of Lithium ion specific power capacity are usually around 100 - 200 Wh/kg, however this post (linked article no longer accessable; see wikipedia) from the electronics stackexchange explores the performance of lithium-air batteries (currently have the highest specific energy of any chemistry) with a value of 1.7kWh/kg for a lithium air battery (Li - O2)

As you can see the absolute worst performing RTG is still several times more energy dense that the best performing chemical battery.

Fuel Cell

As far as chemical energy generation goes, fuel cells are much better, as you can more or less ignore the weight of the fuel cell and only consider the fuel (making the assumption that the mass of fuel is much larger than the cell). Hydrogen fuel cells can reach near 85 - 90% theoretical efficiency from the reaction 2H2 + O2 -> 2H2O (40 - 60% in practice). Even ignoring the efficiency loss (because it's small and I don't want to add the calculation step) we can calculate the specific energy density to be 3.73kWh/kg using the enthalpy of formation of water (the absolute theoretical amount of energy released when hydrogen and oxygen combine)

Even fuel cells at above theoretical maximum performance are not quite as good as one of the worst RTGs (keep in mind we're only considering space applications. RTGs for land use have lots of extra radiation shielding and are very heavy)

  • 7
    $\begingroup$ It is a problem to store hydrogen and oxygen for a fuel cell in space for many years. Using liquid hydrogen and oxygen is possible only if the boiloff rate is as small as the medium consumption by the fuel cell. $\endgroup$ – Uwe May 15 '17 at 19:27
  • 2
    $\begingroup$ @Uwe great point. Fuel cells were used in the apollo missions, but are not a great candidate for satellites. Another big concern is hydrogen embrittlement over time. $\endgroup$ – Aaron May 15 '17 at 19:28
  • 7
    $\begingroup$ Li-air cells would help only if there is air for free. Do the ~1700 Wh/kg include the storage of the necessary oxygen? $\endgroup$ – Uwe May 15 '17 at 19:30
  • 2
    $\begingroup$ As an aside, English plurals do not carry an apostoph. $\endgroup$ – Peter - Reinstate Monica May 15 '17 at 19:32
  • 2
    $\begingroup$ @PeterA.Schneider not according to The Times. (which references the Chicago Manual of Style) $\endgroup$ – Aaron May 15 '17 at 19:35

Simply put, RTGs last a long time. Space probes need a reliable, long-lasting power source, since we can't just change the batteries when they run out. An RTG can run for decades with relatively little reduction in power output, unlike a traditional chemical battery.

  • $\begingroup$ Does traditional chemical battery lose power with time? $\endgroup$ – max May 15 '17 at 14:19
  • 2
    $\begingroup$ @max of course it does. That's what happens when a battery dies. $\endgroup$ – Tristan May 15 '17 at 15:12
  • 1
    $\begingroup$ @Tristan: I think max meant self-discharge, not just normal discharge (from usage i.e. drawing current). Some chemistries have very low self-discharge and it drops with temperature. So I guess it’s a manageable problem. $\endgroup$ – Michael May 15 '17 at 17:37
  • 4
    $\begingroup$ Low temperature would decrease self-discharge but also normal discharge. If a battery is too cold, you could not get the nominal power and energy defined at the nominal temperature. The peak current is also decreased with low temperature. $\endgroup$ – Uwe May 15 '17 at 19:21
  • 2
    $\begingroup$ @max Adding to what Uwe said in the previous comment, you can experience that effect perfectly well right here on Earth. You need two identical, reasonably sturdy flashlights and two sets of batteries, ideally identical. Load one set of batteries into each flashlight, turn both on, put one in the freezer and the other out on the kitchen table. Return after an hour and observe the difference in light output between the two. Measure the battery voltage with a multimeter if you are so inclined. You will likely see a marked difference. $\endgroup$ – user May 16 '17 at 16:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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