When designing an RPS, the isotope itself is only part of the equation. The shape that the fuel comes in has huge impact on the overall capabilities, as does the size of the fuel, the way its clad, which matrix is used, and a host of other considerations. The addition of many different isotopes in the history (and possibly future) of RTGs makes this even more complex.

SNAP Fuels

Let’s begin looking at the way power sources are made by going back to the beginning: the SNAP RTG fuels.

The SNAP-1 was not an RTG at all, because it didn’t use thermoelectric generators. Rather, it used a mercury boiler system (this was the same time as SNAP-2 and -8 were being designed, so in a way this is no surprise). It also used the relatively short half-lived 144Cm in the form of an oxide, beginning a long tradition of using oxides for their chemical stability and thermal resistance for radiosiotope fuel forms.

A good example of a metal early radioisotope fuel can be seen with the SNAP-3A fuel, pictured below.

SNAP-3A Heat Source, image DOE

This was the first RTG to fly on the Transit-4A and -4B satellites, and used 210Po as the fuel. A mix of nested 304 stainless steel cylinders, with a molybdenum or Haynes-25 alloy frustrum, both shielded the payload from braking radiation from the alpha emitter and effectively transferred heat to the thermocouples. A different fuel geometry for a 90Sr fuel element at the time could be seen as part of the SNAP-7D terrestrial RTG below.

SNAP-7D fuel canister, image DOE

The design evolved over time, with each isotope being contained differently. However, over time a multi-layered approach was taken to all fuel types: a chemically and mechanically compatible material is in direct contact with the fuel, followed by a layer designed to mitigate any mechanical damage to the fuel and clad should a catastrophic failure to launch or to orbit occur, follwed by another interface layer to the thermocouples. Other layers can be added depending on the needs of the isotope, like gamma radiation shielding (which is usually only minimal since low gamma flux is a key deciding factor in isotope selection in most cases).

Starting with the SNAP-19, 238Pu became available, changing the requirements for fuel element containment and shielding – by making it easier. Here, the earliest designs used steel/irridium alloys surrounded by Hastelloy steels for containment and preventing radioisotope release in case of a failure of the launch vehicle. Sadly, the first time the SNAP-19 was used showed not only that it could survive, but be reused, when the Atlas launcher for the Nimbus-3 satellite blew up on the ground. When the Nimbus-3B was eventually launched the next year, the same RTG was used on the new spacecraft.

SNAP-27 fuel casing, image DOE

238Pu became an ever-increasingly-popular radioisotope, due to its fairly benign nature, decent power density, and long half-life. The SNAP-27 RTG used microspheres of 238Pu fuel contained in an annulus, with a void for the helium buildup caused by alpha decay. This was the only RTG used on a crewed space mission, and uniquely the only one to be fueled during the mission – the fuel canister and generator were shipped to the Moon separately, then integrated on the Lunar surface. There’s a good picture of Alan Bean fueling the RTG below (via NASA).

If you want to know more about the RTGs used during this time, including source documents on the fuel forms used, check out our SNAP RTG page, available here.


All American RTGs after the SNAP RTG were fueled using 238Pu. There are two distinct phases of this development process: the pre-GPHS RTGs, and the post-GPHS RTGs.

Multi-Hundred Watt RTG (MHW-RTG)

The power supply to the stars, the MHW-RTG is the power source for both Voyager spacecraft. Fueled by 238Pu, this RTG used a silicon-germanium thermocouple for power conversion, and featured smaller-than average fins on the exterior casing due to larger surface area and higher operating temperatures than the TaTe or PbTe thermocouples used on previous RTGs.

The previous 238Pu RTG design, SNAP-27, used an unusual but logical fuel form: microbeads of 238Pu, held in a Hastelloy casing to prevent damage and radiological release into the environment should a failure-to-orbit occur. While this never occurred on any Saturn V launch, one mission did end up bringing a fully fueled RTG back to Earth, re-entering the atmosphere and ending up submerged in the South Pacific. This was Apollo 13, when the LEM and associated systems were used as a lifeboat to bring the three crew back from their disastrous trans-lunar cruise. Due to the fact that the generator and fuel were not mated, and even if they had been the electrical leads were not connected to anything (or available TO be connected), it was useless for the rescue mission, but when the LEM re-entered the atmosphere so did the RTG. No significant radioactive release was ever detected, though, so the transport canister did its job.

That didn’t keep NASA from worrying, or requiring changes to RTG design in the case of a fueled RTG going through a failure to launch or orbit. This means larger fuel elements to allow more clad thickness per fuel element without losing too much system mass. With the increased power supply needed for deep space missions, this demanded a new RTG design.

The new design was called the Multi-Hundred Watt RTG, and the entire program was on a schedule. The Grand Tour needed a power supply for its spacecraft, and this was going to be it.

MHW RTG cutaway diagram showing fuel pellets, image NASA

The 238Pu itself would be covered in two layers of impact protection: one focused on absorbing the energy of the impact itself, the other focused on shearing and other mechanical failures once the fuel cask had impacted.

The fuel was stored in a cask of graphite, coated on the interior with irridium and covered on the exterior with an ablative material. While early RTG designs were designed to burn up in the atmosphere, distributing the fuel as widely as possible to ensure that the effective dose was distributed as widely as possible, times had changed – and with them, the guidelines of responsible RTG use. Now, any radiological release was unacceptable.

However, in tests the fuel pellets didn’t provide as much confidence as they hoped. Failure rates in fuel pellets were significant, including shearing failures in the fuel pellets along weld lines. With the Voyagers waiting, though, the design was completed as a stopgap until a more advanced RTG was available.

Post-explosive test PISA fuel, image NASA

The MHW-RTG has been incredibly successful, continuing to power Voyager 2 (barely, but just enough) as it exited the Solar System. For NASA and JPL, though, improvements could be made. A program to develop a new design for RTG heating units was initiated: the General Purpose Heat Source, which is still in use today.

A page on the MHW-RTG is under construction now, and will be linked here when available.

General Purpose Heat Source (GPHS)

The GPHS is the power supply for every American RTG currently flying other than Voyager. There are two currently used and one flight-qualified radioisotope power system which use this fuel form: the GPHS-RTG (which powers New Horizons, and powered Cassini and Galileo), the MMRTG (which powers Curiosity and will power Mars 2020), and the Advanced Stirling Radioisotope Generator (ASRG), which is proposed as a next-generation power source for future missions. These modules use 238Pu as their fuel. For more information on 238Pu’s production and use, check out the Plutonium-238 page here.

In addition, the GPHS inspired the design of the ESA RTG fuel canister, a collaboration between Lockheed Martin (the contractor responsible for building the fuel cask), the University of Leicester (who make the fuel element), and ESA.

The GPHS is designed with a defense-in-depth strategy, with multiple layers of protection for radiation, mechanical damage, thermal damage, and other characteristics, while ensuring thermophysical and chemical compatibility with both the fuel and the generator itself.

GPHS fuel element in clad, image DOE

The fuel itself is in the form of a domed cylinder, clad in iridium. This fuel form allows for maximum thermal distribution during re-entry, ensuring that thermal failure is minimized. These fuel pellets are stacked with spacers in between, to account for any fuel swelling as well as to provide a more even thermal distribution.

Surrounding the clad fuel are a number of different layers of protection and thermal materials. The first is a graphite impact shell, meant to protect the fuel itself from being released should a failure-to-orbit or launchpad explosion occur. Even though this unit would likely be replaced in the case of a failure like Nimbus, the fuel itself would only likely need a physical examination to verify that it was able to be loaded into a new GPHS module.

These assemblies are called the “fueled GIS,” of which two are used in each GPHS module. The are placed side-by-side horizontally in the GPHS unit within an additional layer of carbon-bonded carbon fiber (CBCF) between the GIS and the half-cube form factor of the GPHS unit itself for additional mechanical protection of the fuel. A cap of the same material as the rest of the aeroshell is then placed over the CBCF cylinder, and a lock screw holds each assembly into place.

The half-cube itself forms an aeroshell, designed to protect the inner modules from the heat of re-entry. A large amount of work has gone into selecting the proper materials for this aeroshell, and the design has continued to be tweaked over the years as new materials have become available, suppliers change their formulas, and thermal modeling has improved, but the overall changes have been minor.

This radioisotope heat source remains the gold standard for RTG design in Western space programs, as can be seen in the 241Am RTG design for ESA. While a failure-to-orbit has never occurred with a GPHS, it is highly unlikely that this would result in radiological release.

Additional information about the development of the GPHS can be found below, and a forthcoming page on the GPHS-RTG – the first type of RTG to use this design – will examine the design as well.

References and Further Reading

General-Purpose Heat Source Development Phase II—Conceptual Designs, Snow et al 1976 https://www.osti.gov/servlets/purl/6569118

Heat Source Component Development Program QUARTERLY REPORT FOR PERIOD JANUARY 1979-MARCH 1979 https://www.osti.gov/servlets/purl/5951012

General-Purpose Heat Source Development: Extended Series Test Program Large Fragment Tests, Henrickson et al https://www.osti.gov/servlets/purl/5857636

General-Purpose Heat Source Safety Verification Test Series: SVT-11 Through SVT-13, George et al 1986 https://www.osti.gov/servlets/purl/5664400

Extending the Useful Life of Radioisotope Thermoelectric Generators through Active Power Control, Raab et al https://www.osti.gov/servlets/purl/1047822


ESA RTG Heating Unit, image University of Leicester (apologies for the poor quality, it’s the best I could find)

As mentioned above, ESA’s RTG is built around a heat source very similar to the GPHS, with only minor changes needed to adapt to the new fuel.

The fuel itself is 70% Am2O3, 30% niobium. This is then clad with an interior protection clad made up of pure niobium. A second clad layer of a 70% ziriconium boride and 30% silicon carbide (by volume) provides additional protection. This forms the outer layer of primary protection for the fuel.

Image University of Leicester/ESA

One or two additional cylinders surround the fuel stack itself, but I’m unable to find out what the materials themselves are, or even verify whether one or two cylinders are used. If we extrapolate from the GPHS, if there ARE two cylinders, the interior one is graphite and the exterior one is carbon bonded carbon fiber (CBCF). Based on the image above, which compares “old” fuel containment with ESA’s design, it appears that the graphite containment has been discarded, and only the CBCF is used as secondary containment within the aeroshell itself. However, this is largely speculation on my part, and should be taken with a healthy grain of salt.

As more details of ESA’s RTG become available, this section will be updated.

References and Further Reading

Space Nuclear Power Systems: Update on Activities and Programmes in the UK, Ambrosi 2015 http://www.unoosa.org/pdf/pres/stsc2015/tech-15E.pdf

AMERICIUM-241 RADIOISOTOPE THERMOELECTRIC GENERATOR DEVELOPMENT FOR SPACE APPLICATIONS, Ambrosi et al 2013 https://inis.iaea.org/collection/NCLCollectionStore/_Public/45/066/45066049.pdf