Radioisotope Power Sources

We spend most of our time at Beyond NERVA looking at nuclear reactors and their support systems, but this isn’t the way that nuclear power has been used in the majority of cases when it comes to spaceflight. Instead, today we use radioisotope power sources, usually plutonium-238, that produce decay heat which is then used either AS heat to keep electronics and sensors at the correct temperature (known technically as a radioisotope heater), or as an electrical power source, but often as both.

There are two ways that nuclear power can be extracted without using a reactor: through radioactive decay and through spontaneous fission. There’s a few different ways to use either option, but we’re going to focus on the radioactive decay part here, and specifically at radioisotope power sources (RPS) that use radioactive decay to produce heat. This heat is then used to keep spacecraft components warm in space, or to produce electricity.

The first option is to just use the heat to keep components warm. The most common application of this is the US “General Purpose Heat Source,” or GPHS. This is confusing, because there’s an American COMPONENT called a GPHS as well (and, yes, it’s the same when you unpack it). This is usually a small piece of a radiation-shielded radioisotope, with heat pipes connected to the outside to carry heat to the (mostly) electronics and bearings that need a certain minimum temperature to operate.

The other popular option is to produce electricity using that heat. While there are many options on the physical mechanism that allows this to happen, one has truly overwhelmed the others when it comes to real-world application. This is the venerable RTG, or Radioisotope Thermoelectric Generator. These systems have flown to every planet in the Solar System but Mercury, landed on one of them (Mars), and powered the manned Apollo moon missions as well. If you want a simple system, it’s hard to go simpler: a plug of radioactive material generates heat during radioactive decay, and that heat is converted to electricity (usually with thermoelectric convertors). They are reliable, almost impossible to break (just don’t trip over the power cables!) and incredibly well tested, but these are very low-power systems, and while the waste heat that they produce is often budgeted very carefully for electronics and sensor thermal management, it’s something that just has to be dealt with in most cases.

However, RTG power outputs are low. Looking at the various systems that have been flown, and ones that are currently worked on, you see that all of them are measured in watts.

Table via Wikipedia, 2017

There are ways around this, however. Heat engines, such as Stirling, Rankine, and Brayton engines, have all been proposed for isotope-powered systems, both historically and recently with NASA’s ASRG, and a mention of possible ESA expansion into a Stirling variant of their new RTG as well. These offer much higher efficiencies, at the cost of far more complex systems, which are more likely to break, and as such have never flown.

The final option, and one that was once far more popular than it is today, was to use the heat of the decay to drive a thermal engine. This is commonly known as a Radioisotope Thruster (RIT), but is colloquially known as a Poodle thruster after the best known program looking in to them. (the smaller cousin to the Rover Program going on at the time). These are reasonable thrust level, moderate specific impulse engines, which are either limited by the short useable half-life for higher-operating-temperature fuels, or by the lower operating temperatures of longer-lived but less active fuels. This doesn’t mean that they’re a bad system by any stretch of the imagination: often they’re able to double the isp and reduce mass of comparable resistojet thermal thrusters.

Radioisotope Fuel Elements

We’re going to be talking a lot about radioisotopes, which make up the fuel for an RTG, so let’s go ahead and define the term off the bat. First, the term “isotope,” which just means a particular atomic mass of a particular element. Elements are defined by the number of protons in the nucleus, but the number of neutrons varies. Some isotopes have too much energy, and need to reduce the amount by going through radioactive decay, releasing energy in the form of a helium nucleus (alpha radiation), an electron or positron (beta radiation, called “beta-minus and beta-plus respectively) – with the side effect of a neutron effectively becoming a proton, a high energy photon (gamma radiation), or in very rare cases either a neutron or through spontaneous fission.

A number of considerations come into play for any nuclear fuel element, but for radioisotopes many of these become far more complex. One of the biggest is that your fuel element’s chemical composition changes over time, which needs to be accounted for as far as the overall stability of the fuel element itself. Another is the half-life of the fuel being used (which determines the useful life), as well as the power density of the isotope (determining how much fuel mass is needed), as well as the type of radioactive decay the isotope goes through. For more information about selecting the radioisotope, check out our Radioisotope Selection page, including information about some of the many options for which isotope has been – or can be – used for astronuclear power systems.

After isotope selection, a fuel element is developed. This has to perform many of the same functions as fission fuel, but without the actual process of fission, or significant neutron radiation. However, the combination of high-temperature operation, a greater or lesser radiation environment (with the same analytic skills required in many areas), the balance of power generated with thermal transport, and other considerationsare common to both. As with fission fuel, oxides are the most common, although CERMET fuel elements have been proposed. Also like fission fuel, the fuel element is then clad in a material to chemically protect it, as well as preventing accidental release.

Learn more about these systems on our Radioisotope Fuel Form and Containment page:

Finally, the fuel element is placed in a structure to use the heat generated. This not only integrates the thermocouple to the fuel element, but it also provides radiation shielding, impact protection, launch protection, and, in a worst-case scenario, lithobraking protection (i.e. it won’t crack open if the launch fails). This can either be a self-contained subsystem, such as the General Purpose Heat Source (GPHS) used by the US designs, or it can be integral to the spacecraft itself, but it’s far more common to have a GPHS-style architecture on modern RTGs.



For more information on radioisotope selection, and the options used in the past, check out our Radioisotope Fuel Selection page, here!

Fuel Element Design

In many ways, the design of a radioisotope fuel element is largely similar to that of a fuel element, with some significant differences – both in the positive and the negative. Both have to deal with power balancing and thermal distribution throughout the fuel element, with both active and passive components to the material, both have to handle high temperature operation in a radioactive environment, both have to balance power production with thermal conductivity, and both have to address the changing chemical structure of the fuel element over time. However, there are a number of major differences: the fission element has to deal with far higher radiation flux, including an intense neutron flux, deal with far higher thermal cycling, deal with far higher average power distribution (typically), and undergoes extreme changes in chemical composition; radioisotope fuels, on the other hand, have to be as self-shielding as possible against all undesired forms of radiation, must be easily handled and integrated at the launch facility, and often must perform constant duties for much longer than a fission fuel element would ever have to.

Just as with fission elements, radioisotope fuel elements are first made using the chosen isotope and high-purity, sometimes isotopically enriched, materials that are used to form a stable chemical structure with as close to the desired thermal, chemical, and radiological properties, balancing operating temperature, power density, thermal conductivity, radiation shielding requirements, and other characteristics. Often this is an oxide, but oxide fuel distibuted through a refractory metal matrix – CERMET – has been proposed for RTG fuel before.

Finally, usually the fuel element is clad in a material to make chemical compatibility easier, increase radiation shielding, and protect the fuel element in the case of a failure-to-orbit. The material is chosen based on its thermal expansion properties, chemical compatibility with both the fuel and everything else it’ll come in contact with, and its radiation shielding capabilities, among others

The Radioisotope Fuels page covers these considerations for the different isotopes more in depth, and over time each isotope will have its own page with even more detailed information as well.

Radioisotope Thermal Rockets: The Resistojet’s Sleeker Cousin

Three unit Poodle thruster, image DOE

Radioisotope Thrusters, or RITs, use the decay heat of the fuel to heat a propellant for a thermal thruster. These systems are at the half-way point between a nuclear thermal rocket and a resistojet thruster, which we examined very briefly in our look at electric propulsion [insert link].

These systems are useful in a couple ways: first, as the interplanetary equivalent of a JATO (Jet-Assisted Take-Off) bottle, and second as a high specific power monopropellant reaction control thruster, similar to an upgraded cold gas or resistojet thruster. Which way they’re most useful depends highly on the isotope used, the propellant used, and the mission profile they are applied to.

The most well known of these systems is the result of Project Poodle, a program contemporaneous with Project Rover (hence the name). This thruster used 210Po fuel (one of the few to use this short-lived isotope for space missions), and hydrogen propellant. It produced 1.1 N of thrust with a specific impulse of 710s.

DART turuster, notice the four nozzles on the right hand side. Image DOE

Another design used a longer-lived isotope, 238Pu, the same that has powered missions familiar to most readers as the power supplies for everything from Apollo’s lunar science experiments to Voyager to Curiosity. This one used hydrazine, which decomposes in the heat of the fuel element, producing thrust the same way that a resistojet does (the original design, from GE, actually WAS an adapted resistojet).This would be more useful as a reaction control thruster, with the longer half-life and greater thrust offered from the hydrazine than the hydrogen.

This is the subject of a blog post coming out shortly after this website goes live (1-2 weeks), so for now we’ll leave it here, but expect more information in the near future.

Power Conversion Systems

The vast majority of radioisotope power sources convert the heat of radioactive decay into electricity, often using the waste heat to keep sensitive instruments warm as well. These Radioisotope Power Units, or RPUs, come in many different varieties, with the most common being the radioisotope thermoelectric generator, or RTG. This uses a thermocouple to convert the heat into electricity in a very robust, solid-state system.

However, low power conversion efficiencies, increasing power requirements, and improved manufacturing techniques, as well as an incredibly successful test campaign, have opened up the door for the use of heat engines in RTGs with the Advanced Stirling Radioisotope Generator. Other proposals for both Rankine and Brayton cycles have been proposed as well, offering much higher efficiencies, at the cost of vastly increased complexity in some cases. 

Radioisotope Thermoelectric Generators

MMRTG cutaway, image NASA

RTGs convert heat into electricity through thermoelectric converters. The thermoelectric effect occurs when there’s a difference in temperature across the junction of two different metals or semiconductors. How efficiently this is done depends on a number of things, including the microstructure and crystalline structure of the materials being used for each element of the converter, meaning that changing the alloy of the base materials used can have a significant effect on the performance of the converter.

This application was actually widely in use already as a type of sensor called a “thermocouple,” which sends a voltage based on its temperature within a certain range of environments and is one of the most common forms of temperature monitoring in many fields.

Designers of early astronuclear systems, such as the designers of the SNAP-3 RTG (which flew for the first time in 1961) scaled this concept, turned it inside out, and placed it on a spacecraft. Many different types of material combinations have been experimented with over the years, mostly based around lead and tellurium, PbTe. Thin strips were alternated around the radius of the fuel canister, with heat being provided by conduction and radiation using wide radiator fins. Nowadays, the conversion unit is an array of thermocouples in a cylindrical form around the fuel element itself.

Today many designs still use lead telluride (PbTe) doped with a special material that was invented in the 1960s: TAGS-85. In TAGS, the tellurium used in the converter. is carefully doped with silver (Ag) and antimony (Sb), hence TAGS. The most commonly used version of TAGS, TAGS-85, uses (PbTe)85(AgSbTe2)15. This type of material was first used in the SNAP-19B RTG, for the Nimbus-B satellite which failed to launch. The fuel was recovered, and placed in the Nimbus-3 satellite, which launched April 3, 1969. It continues to be used to this day on the Mars Science Laboratory, Curiosity, and will fly on the Mars 2020 rover as well.

Another, possibly more common option with about as long a history is silicon-germanium, a better option at higher temperatures. This has been used in the SNAP-27 RTG, which was part of the Apollo astronaut’s Apollo Lunar Science Experiments Package, and several of these units remain on the moon. Another frequent use for this type of thermocouple is in the GPHS-RTG, which powered multiple Pioneer spacecraft as well as Voyagers 1 and 2.

ESA is looking at a variety of options as well for their 241Am fueled RTG. Their preferred material currently is bismuth telluride, which is efficient at the desired temperatures (241Am has a lower fuel temperature when compared to 238Pu), as well as being commercially available. However, they looked at a number of other options as well during design.

Another promising option is a class of materials called skutterudites. These materials are large structures, made up of materials such as cobalt, rhodium, antimony, and others. These materials have low thermal conductivity and high mobility, exactly the things you look for in a thermocouple material.

There are many other thermoelectric possibilities, and over time I hope to add to the list. I am currently in the process of putting together a page on thermoelectric generators used or proposed for space missions currently, and will update with a link when that’s available.

Radioisotope Heat Engine Generators

The last choice available that harnesses heat to produce electricity is using a heat engine, as is usually done in a reactor. These use mechanical systems to convert heat to electricity. These are common types of power conversion systems on Earth, as well as one very old one which is making a strong comeback.


The Stirling engine has once again found its way into the limelight, and its inherent simplicity and high potential efficiency (for a Carnot cycle-dependent conversion system) make it attractive for use in spacecraft. Because of this, NASA is looking at its first non-materials-based power conversion system for a radioisotope power source (RPS): the Advanced Stirling Radioisotope Generator and ESA’s Stirling radioisotope power unit. ESA is also looking into using Stirling engines for their new 241Am fueled RHU, where the increased efficiency is able to extend the power output of their lower specific power fuel.

Once again, these are subjects that we’re going to be covering in the coming months, and a page for each concept will be forthcoming as well.

Radioisotope Power Sources

SNAP: The Beginning of an Era

SNAP-19 RTG which powered Pioneer 10 &11, Viking landers 1 & 2, image NASA

The Space Nuclear Applications Program, or SNAP, was a large program looking at nuclear power sources (and related technologies), mainly in the 1950’s through 1970’s. Even numbered SNAP programs were nuclear reactors, and odd numbers were for radioisotope power sources. While we tend to focus on the reactors (SNAP-2, SNAP-8, SNAP-10 and SNAP-50, links to the pages available by clicking on the names), the RTGs developed during the program were far more commonly used.

These power systems were not only designed for space, but terrestrial and nautical use as well. Various different designs have been used everywhere from underwater to the very ragged edges of our solar system, and landed on the Moon and Mars.

For additional information on these systems, as well as their use in solar system exploration, check out the SNAP RTG page, available here.

MHW-RTG: Humanity’s Power Supply in Interstellar Space

Voyager MHW-RTG

The Multi-Hundred Watt RTG (MHW-RTG) is the power supply for the Voyager spacecraft. It used the same silicon-germanium thermocouples as the SNAP-27 ALSEP RTGs that were used during Apollo on the surface of the Moon paired with 238Pu fuel. Unlike in both earlier and later RTGs, the MHW-RTG used spherical fuel elements contained in a cylindrical housing. Each fuel element was individually clad in irridium alloy.

Each Voyager uses three RTGs to provide power, and despite being launched in the 1970s and being outside the Solar System itself, is still able to keep the spacecraft limping along and sending back useful data. Two other satellites used this technology, but it was soon phased out for the similar mission requirement GPHS-RTG, which improved fuel safety and handling characteristics, as well as returned the fuel form to a cylinder.

A page for the MHW-RTG is coming, and will be linked here.

GPHS-RTG: Exploring the Giants and the Dwarves

GPHS-RTG Cutaway diagram, image DOE/NASA/JPL

The GPHS-RTG takes another heritage thermocouple material, in this case the silicon-germanium thermocouples used in the MHW-RTG, and the SNAP-27 before it. These have been used on many missions, most notably the Cassini, Galileo, and New Horizons spacecraft. To my knowledge there are no currently funded plans to fly another GPHS-RTG, but it remains a popular benchmark power supply, especially for outer solar system missions (e.g. “with x number of GPHS-RTGs”).

More information on the GPHS-RTG is on a dedicated page coming soon, and will continue to be added to over time.

MMRTG: NASA’s Workhorse on Mars

MMRTG on Curiosity Rover, image NASA/JPL

The Multi-Mission RTG, or MMRTG, is best known as the power supply for the Curiosity rover on Mars, and will also power the Mars 2020 rover. This RTG uses the same General Purpose Heat Source fuel element as all modern American RTGs use, and the same Pb-TAGS-85 thermocouple as was used on Viking and Pioneer 10’s SNAP-19, with some doping changes to improve the efficiency.

More information on the MMRTG can be found on its own page, which will be up (but not complete) shortly.

More on the GPHS RTG can be found on this page, which will continue to be filled in over time.

NASA’s Next Generation RTG

NASA and JPL both have been investigating the next step in RTG systems, with many options on the table since the last major upgrade to thermoelectric convertors. The advances of the 1990s and 2000s in thermoelectric materials have not been applied to RTGs yet, and these advances have continued since so there’s even more options available.

NASA is working on developing their next solid-state radioisotope power supply, with bidding actually underway at the time of this writing (5/19) for the next thermocouple material to be used in the new generation of RTGs.

Proposals have been submitted, and a communications blackout of all NASA and JPL personnel involved in the program has been implemented. This means only one thing: goodies are coming soon!

Information about the different options available, and the proposals that have been made in the past, will be available on the Next Gen RTG page, which is currently under construction. Additionally, the designs not selected, and additional materials options not mature enough for the selection criteria, will be covered on the Thermoelectric Generators page, which is also under construction.

ASRG: A New Breed of Radioisotope Power Supply

ASRG Cutaway, Wong NASA 2015
ASRG Cutaway, ASC = Advanced Stirling Convertor, GPHS = General Purpose Heat Source, via NASA

A newly tested (but not proposed) idea in the RTG world is to move away from materials-based conversion systems and on to heat engines, in particular the Stirling engine. With modern materials, tooling, and design, the few moving parts in a Stirling engine make it an incredibly attractive option. The latest design to come out of NASA for a radioisotope power source is the Advanced Sterling Radioisotope Generator (ASRG). NASA has tested the Sunpower Advanced Stirling Convertor for well over 30,000 hours now as part of the ASRG Engineering Unit (ASRU-EU). With this setup, NASA hopes to achieve >35% efficiency, with about 140 W of electrical power being produced, and the subsystems have all been thoroughly tested. Despite this, the program was cancelled for budgetary reasons after no mission that needed the ASRG was selected. As of 2015 the first flight article isn’t currently expected to fly before 2028, assuming there’s a mission for it by then. Until then the workhorse MMRTG will continue to be used.

Advanced RPSs in the Past, and the Future

Not every design for an RPS made it off the drawing boards. While the vast majority of US designs have used 238Pu for spacecraft missions these days, the broad range of radioisotopes available as sources show that many more designs are possible.

One option that was proposed, and subsequently forgotten, was for a thermionic RTG. This seems to have been a single proposal, and US thermionic technology was rudimentary at best. While it would be interesting to look at a thermionic conversion system using modern design, it is unlikely that the small efficiency gains over solid-state conversion systems will be a better choice than the far more efficient Stirling generator or other heat engine. The paper can be found here:

Some early radioisotope power sources used Rankine cycle conversion systems on paper, but materials challenges and development time (and cost) led mission planners and program funding to shift to solid state conversion systems. These included the SNAP-1 RTG, the first of the SNAP designs proposed, which used a very similar conversion system to the early SNAP-2 reactor, a mercury Rankine system driven by Cerium-144. Others have been proposed over the years as well, and recent advances in organic Rankine cycles offer promise for RPS as well, especially at lower temperatures and specific powers.

One design which was quite popular in the early days of the Modular Space Station planning (the great-grandfather on the American side of the ISS) used a helium Brayton cycle with pebbles of 230Pu oxide for fuel. Other designs have also been proposed, including one in recent years from the University of New Mexico, but despite their promise, the aversion to moving parts in spacecraft has led to their never being flown.

There’s even a proposal for an AMTEC, or Alkali Metal Thermal Electric Convertor, RTG! I’m going to hold that one close to the chest before I write it up, but there will be a page available once it’s done.

As both materials and designs improved, these concepts were re-proposed, but none have ever flown. In the coming months, I’ll be covering these designs in the blog, and will add pages for each, and summaries (with links) here.

These designs would each have their own advantages and challenges, and unique – and useful – plans wait to be either dusted off and updated. Pages for each, as well as for each type of power conversion system, will be appearing over the coming months.


From a practical engineering point of view, most RPSs are, and will continue to be, low-power systems. They have been used far more extensively in Russia than in the US, for a wide variety of applications and using a wide variety of different radioisotopes. There are a lot of advantages to their simplicity and ruggedness, and they will continue to fill a vital role in space missions for the indefinite future. However, the practical limitations on them also mean that there’s times they just won’t work.

This doesn’t make them useless by any stretch. More than anything, the limitations on radioactive material production prevent them from being a common power source due to their reliability, predictability, and (mostly) environmental independence to power the most distant missions humanity has ever attempted, and may continue to do so long into the future, not only for electrical power but for thrust as well.

Further Reading and References

Mission Proposals

Radioisotope power systems and their applications for solar system exploration, Balint 2004

Enabling exploration with small radioisotope power systems, Abelson et al 2004

Small RPS Enabled Deployable Mini-Payload Missions STAIF 2005

Enabling future low-cost small mission concepts, Lee 2014

Enabling future low-cost small spacecraft mission concepts using small radioisotope power systems, Lee 2014

Outer planet spacecraft temperature testing and analysis, Hoffman et al 2002

Advanced Radioisotope Power Sources for Future Deep Space Missions, Nilsen, E. 2000


Searching for subsurface lunar water ice using a small RPS-powered rover, Randolph, J et al 2005


Nuclear power options for Mars polar robotic outpost, Elliott et al 2004

Nuclear systems for Mars exploration, Balint 2004

The Mars Environmental Survey (MESUR) Network and Pathfinder Missions, McNamee 1993

Mechanically Pumped Fluid Loop (MPFL) technologies for thermal control of future Mars Rovers, Birur et al 2006

Jupiter System

Innovative Balloon Buoyancy Techniques for Atmospheric Exploration Jones, J 2000 (Titan RTG)

Titan Inflatable Aerovehicle/Rover/Boat Jones, J 2000

Titan in situ exploration concepts at JPL, Elliott 2008

Advanced radioisotope power system enabled Titan Rover concept with inflatable wheels. Balint et al 2006

Saturn System

A radioisotope powered cryobot for penetrating the Europan ice shell, Zimmerman et al 2001

Europa orbiter mission concept, Park, Y. H. 2002

Exploring Europa with a surface lander powered by a small radioisotope power system, Abelson 2005

Assessment of alternative Europa mission architectures, Langmaier et al 2008

General Reference and Mission Design

Big things come in small packages: mission concepts potentially enabled by small radioisotope power systems, Abelson 2005
Part 1
Part 2

The year in review : power systems in space exploration, Timmerman 2005

Radioisotope power systems reference book for mission designers and planners Lee, Young; Bairstow, Brian 2015



Thermally Cascaded Thermoelectric Generator 1970

Thermally Cascaded Thermoelectric Generator Final Report 1970

Thermoelectric Generators for Deep Space Application, 1971

(paywall) Safe radioisotope thermoelectric generators and heat sources for space applications, O’Brien et al 2008



U.S. Space Radioisotope Power Systems and Applications: Past, Present and Future Robert L. Cataldo and Gary L. Bennett


NASA’s Radioisotope Power Systems Planning and Potential Future Systems Overview 2016

Fuel Element Manufacture


Environmental and Radiological Impacts

Compatibility and Shielding Analysis of Science Instruments in Spacecraft Containing a Radioisotope Thermoelectric Generator 1970


Monte Carlo Radiation Analysis of a Spacecraft Radioisotope Power System, Wallace 1994

Safety Monitoring System for Radioisotope Thermoelectric Generators 1972


Space nuclear power public and stakeholder risk communication, Dawson 2005

Aging and End of Life Considerations

Aging Effects of US Space Nuclear Systems in Orbit Bartam et al 1982



Decommissioning of the Special Mettalurgical Building at Mound Laboratory, Harris et al

Various Programs Worldwide: Status Reports, Final Reports, and other Errata


Technical Progress Reports: RPS Materials Production and Technology Program


January through March 2000

April through June 2000

July through September 2000


September 2000 through March 2001

OCTOBER 1, 2001 THROUGH MARCH 31, 2002







October 1, 2007 through September 30, 2008



October 1, 2010 Through September 30, 2011