The Pioneer Anomaly: 30 Years of Mystery, One Elegant Solution

Pioneer 10 and Pioneer 11 were the first spacecraft to head for the outer planets, and they finished their planetary missions in the 1970s. Then, as they coasted out of the solar system, radio tracking showed that both were being pulled toward the Sun by about 0.9 billionths of a meter per second squared more than any known force could explain. Physicists proposed dark matter, modified gravity, and new fields. A thermal model published in 2012 pointed to something plainer: heat leaking unevenly from the spacecraft itself.

What Pioneer 10 and 11 did before they became a puzzle

Pioneer 10 launched from Cape Canaveral on the evening of March 2, 1972 local time (early March 3 in UTC). It was the first spacecraft to cross the main asteroid belt, entering on July 15, 1972, and it passed Jupiter in early December 1973. Pioneer 11 launched on April 5, 1973 local time (April 6 in UTC), flew past Jupiter in December 1974, and became the first spacecraft to visit Saturn, passing about 21,000 kilometers from the planet on September 1, 1979.

After those encounters the probes simply kept going. NASA's last signal from Pioneer 10 came on January 23, 2003, from about 12.23 billion kilometers away, a trip that took the radio signal 11 hours and 20 minutes. Pioneer 11's last engineering data arrived on November 24, 1995, according to NASA. Both spacecraft are now silent, but the tracking records they left behind are the reason this story exists. NASA's Pioneer 10 mission page lists these milestones.

How a tiny push shows up in radio tracking

Deep-space tracking works through the Doppler effect. The Deep Space Network sends a radio signal to the spacecraft, which returns it on a related frequency. If the spacecraft's speed along the line of sight changes, the returned frequency shifts, and the shift can be measured with extraordinary precision. Navigators fit a model of every known force to the measured shifts, and whatever the model cannot account for is called a residual.

In the Pioneer data, the residual was a slow, steady frequency drift of about 6 × 10⁻⁹ hertz per second, according to the 2010 Living Reviews in Relativity paper by Turyshev and Toth. That drift is what you would see if each spacecraft were decelerating, with an acceleration pointing roughly toward the Sun, by (8.74 ± 1.33) × 10⁻¹⁰ meters per second squared. That value comes from the definitive analysis by John Anderson and colleagues, published in Physical Review D in 2002, which used about 11.5 years of Pioneer 10 data (January 3, 1987 to July 22, 1998) and about 3.7 years of Pioneer 11 data (January 5, 1987 to October 1, 1990), as listed in the paper's abstract. The authors examined effects both external and internal to the spacecraft and found no systematic origin at the time.

An independent check came from C. B. Markwardt of NASA's Goddard Space Flight Center, whose 2002 analysis of Pioneer 10 data, titled "Independent Confirmation of the Pioneer 10 Anomalous Acceleration," reported (8.60 ± 1.34) × 10⁻¹⁰ m/s². Several groups with separate software finding the same residual made a simple navigation error unlikely.

How small is 8.74 × 10⁻¹⁰ m/s²?

The number is hard to feel, so here are three comparisons. First, Earth's surface gravity is about 9.8 m/s², so the anomaly is roughly 11 billion times weaker. Second, the Sun's own pull on a spacecraft is far larger at Pioneer's distances. Using the Sun's gravitational parameter (about 1.327 × 10²⁰ m³/s²) and Newton's inverse-square law, the Sun's pull is:

Distance from Sun	Solar gravity (m/s²)	Anomaly as share of solar gravity
20 AU	About 1.5 × 10⁻⁵	About 0.006%
40 AU	About 3.7 × 10⁻⁶	About 0.024%
70 AU	About 1.2 × 10⁻⁶	About 0.07%

(One astronomical unit, or AU, is the Earth-Sun distance, about 150 million kilometers. The values are straightforward calculations from the figures above, not measurements.)

Third, here is what a constant acceleration of that size does over time. Navigators care about position, not just speed, and position errors accumulate with the square of time.

Elapsed time	Speed change	Position offset
1 year	About 0.03 m/s (0.1 km/h)	About 435 km
10 years	About 0.28 m/s (1 km/h)	About 43,500 km
30 years	About 0.83 m/s (3 km/h)	About 392,000 km

These come from simple kinematics, assuming the acceleration stays constant, so they are illustrations rather than reported residuals. The first row matches the commonly quoted statement that a Pioneer position predicted a year ahead came out roughly 400 kilometers off. The second row matches the "about 1 km/h per decade" slowdown often cited. Against distances of billions of kilometers, those errors are tiny, which is why the anomaly took decades of careful accounting to establish.

Why the Pioneers were the right spacecraft for the job

Most deep-space probes are poor laboratories for tiny forces, because their own maneuvers add noise. The Pioneers were spin-stabilized: they rotated at about 4.8 revolutions per minute around an axis pointing at Earth, which kept the high-gain antenna aimed correctly without constant thruster firing. A spinning top stays upright for the same reason, with its rotation resisting small nudges.

According to the Wikipedia summary of the anomaly, after the planetary encounters the Pioneers used thrusters only for semiannual conical scanning maneuvers. As Astronomy magazine notes, the Voyagers, which are not spin-stabilized and rely more on thrusters, cannot be tracked to this level of precision. The anomaly itself sits at the 10⁻¹⁰ m/s² scale. For more on how navigators handle spacecraft far from home, see how spacecraft navigate in deep space and the story of Voyager 1 and 2.

The spacecraft's hardware matters for the solution too. Each Pioneer carried four SNAP-19 radioisotope thermoelectric generators (RTGs), power sources that turn the heat of decaying plutonium-238 into electricity. On Pioneer 10 they produced about 155 watts of electrical power at launch, about 140 watts at Jupiter, and about 65 watts by 2001, according to the Pioneer 10 article on Wikipedia. The RTGs sat on two trusses, and a 2.74-meter parabolic dish dominated the spacecraft's outline. Remember those numbers; they come back later.

First reactions: real, or an artifact?

The anomaly was first noticed as early as 1980 but not seriously investigated until 1994, according to Wikipedia's Pioneer anomaly article. Anderson and colleagues published their first detailed paper in 1998 in Physical Review Letters, reporting a Sunward acceleration of roughly 8 × 10⁻¹⁰ m/s², and refined it to the (8.74 ± 1.33) × 10⁻¹⁰ m/s² value in 2002.

The team and its critics checked external effects, such as solar radiation pressure and the pull of distant matter, and internal ones, such as leaking gas and radio-system quirks. Solar radiation pressure, the force sunlight exerts on a spacecraft, is worth a note because it behaves differently from the anomaly: it weakens with the square of distance, while the anomaly looked roughly constant out to 70 AU. A related physics story is told in why solar sails can move without fuel.

Candidate explanations and how each fared

Because the effect looked constant and directed at the Sun, theorists proposed a wide range of causes. The table summarizes the main families and why most did not survive.

Proposed explanation	Idea	Problem
Navigation or modeling error	Wrong planetary positions or software bug	Multiple independent analyses found the same residual
Solar radiation pressure	Sunlight pushing on the craft	Falls off with distance squared; the anomaly did not
Fuel leaks or thruster gas	Gas escaping from the hydrazine system	Hard to produce a steady, Sunward force on both craft
Dark matter near the Sun	A halo adds extra pull	Would also disturb planetary orbits
Modified gravity (for example MOND)	Gravity behaves differently at low accelerations	Solar-system accelerations are far above the threshold
New scalar or vector fields	Additional long-range force	Constrained by planetary data; no independent evidence
Anisotropic heat emission	Uneven thermal radiation produces recoil	Fits in size and direction; confirmed by detailed models

Why MOND and dark matter did not fit

Modified Newtonian Dynamics, or MOND, was proposed by Mordehai Milgrom in 1983 to explain why stars in the outer parts of galaxies orbit faster than visible matter should allow. Its key number is a threshold acceleration of about 1.2 × 10⁻¹⁰ m/s² (a value fitted to galaxy rotation curves in 1991), according to the Wikipedia article on MOND. That is the same order as the Pioneer anomaly, which made the link tempting. The trouble is that MOND modifies gravity only where accelerations fall below that threshold. At Pioneer's distances, the Sun's pull is about 10,000 to 100,000 times larger (see the table above), so MOND's correction there is negligible. In addition, the effect on a spacecraft would not take the shape of a constant Sunward push.

The dark matter idea fails for a different reason. A halo dense enough to tug on the Pioneers would also tug on Uranus, Neptune, and Saturn. E. M. Standish's analysis of planetary ephemerides, presented in the Proceedings of the International Astronomical Union, found that most proposed forms of the anomaly, including a constant acceleration toward the Sun, would be unacceptable for the planets. Some analysts have argued that the outer-planet data are not precise enough to settle the question. For the broader dark matter picture, see dark matter and dark energy explained.

The numerical coincidences

The anomaly sits near two other numbers that invite speculation. One is MOND's threshold acceleration. The other is the product of the speed of light and the Hubble constant, which is (6.59 ± 0.07) × 10⁻¹⁰ m/s² according to Wikipedia's Pioneer anomaly article. The Hubble constant measures how fast the universe expands; our explainer on Hubble's law covers the idea. Because the figures are close, a few authors proposed that cosmic expansion might somehow leak into local dynamics. The same article says physicists consider the match a coincidence. Numbers in the same order of magnitude are common in physics, and without a mechanism, a coincidence is only a coincidence.

The heat hypothesis

The idea that heat could be responsible was raised early. In 1999, J. I. Katz suggested in Physical Review Letters that the effect might be the recoil of RTG waste heat scattered by the back of the high-gain antenna. Anisotropic means unequal in different directions. A photon carries momentum, so a spacecraft that radiates more heat in one direction than the opposite recoils the other way, the way a garden hose pushes back on the person holding it.

The idea is plausible because of how much heat the Pioneers produced. Turyshev and colleagues' 2012 paper puts the RTGs' heat output at about 2,578 watts in mid-1972, and the Pioneer 10 electrical output at launch was about 155 watts. That implies an efficiency near 6 percent, so roughly 94 percent of the RTG output was waste heat. Electronics inside the spacecraft added more heat as they used up the electrical power.

A worked example shows how little of that heat it takes. Force equals mass times acceleration. Using Pioneer 10's launch mass of about 258 kilograms (per NASA) and the anomaly of 8.74 × 10⁻¹⁰ m/s², the force is about 2.3 × 10⁻⁷ newtons. A beam of light with power P exerts a force of P divided by the speed of light, so the force corresponds to about 68 watts of perfectly directed light. Against thousands of watts of thermal power at launch, that is a few percent. The question was whether Pioneer's geometry could send a few percent of its heat in one direction.

It could, in principle. The RTGs and the electronics compartment sat behind the large dish, and heat that hits and reflects from the dish and the spacecraft body does not leave symmetrically. Those effects were hard to quantify from a sketch, which is why the question stayed open until someone built a detailed model of the real spacecraft.

Rescuing the data

Building that model required data that was at risk of disappearing. Pioneer tracking and telemetry records existed on paper and on older magnetic tape formats, and parts of the project record were stored carelessly. According to the Astronomy magazine article, some materials were "literally in boxes under stairwells and in dumpsters on their way to being destroyed." Recovering them meant finding the sources, restoring hardware that could read the tapes, and converting the contents to modern formats. Cost and manpower were the main obstacles.

Slava Turyshev and collaborators led the effort, and The Planetary Society helped raise funds. Wikipedia records a Planetary Society update titled "Data Saved!" dated June 1, 2006. The recovered material fed the 165-page review that Turyshev and Viktor Toth published in 2010 in Living Reviews in Relativity, which documents the spacecraft's design, operations, and data formats. The archive turned a 1970s mission into a dataset that could answer questions nobody had asked in the 1970s.

The 2012 thermal model: what Turyshev's team did

The decisive paper was "Support for the Thermal Origin of the Pioneer Anomaly," by Turyshev, Toth, Gary Kinsella, Siu-Chun Lee, Shing Lok, and Jordan Ellis, published in Physical Review Letters 108, 241101 (2012). The authors' full text describes the method.

They built a finite-element thermal model: the spacecraft was divided into small surface elements and nodes, and the equations for heat conduction and radiation were solved across all of them. The model used about 3,300 surface elements, 3,700 nodes, and 8,700 linear conductors. There were no CAD files for a 1970s spacecraft, so the team constructed the geometry by hand from blueprints and project documents, including the behavior of the louvers (temperature-controlled shutters). They ran it in commercial thermal-analysis software and used flight telemetry as boundary conditions, then evaluated the model at eleven distances from 3 to 80 AU.

Telemetry mattered because the heat budget changed over the mission. The plutonium-238 decays with a half-life of 87.74 years, so the RTG heat fell slowly, and the electrical power available to the electronics fell with it. The paper's power table for Pioneer 10 shows the RTG-generated power dropping from 148.5 watts at 3 AU to 67.2 watts at 70 AU. A recoil force that falls with the power budget would produce an anomaly that shrinks with time, and the data suggested that it did.

The team then compared two things. One was the thermal recoil predicted by the model. The other was a fit of the same kind of recoil model to the Doppler data. The comparison, from the paper:

Quantity	Thermal model (telemetry)	Fit to Doppler data
Efficiency factor for RTG heat	0.0104	0.0144
Efficiency factor for electronics heat	0.406	0.480
Decay time constant (from 1980)	36.9 ± 6.7 years	28.8 ± 2.0 years
Acceleration at 1980 (in 10⁻¹⁰ m/s²)	7.4 ± 2.5	10.1 ± 1.0

The efficiency factors say what fraction of each source's heat becomes recoil along the spin axis. The thermal estimate came to about 80 percent of the Doppler-derived magnitude, with roughly 35 percent of the anomaly traced to the RTGs and 45 percent to the electronics. The paper reports that the error ellipses of the two estimates overlap, so there is no statistically significant difference between them. The conclusion, in the authors' wording, was that once thermal recoil is included, "no anomalous acceleration remains."

How settled is it?

Most researchers regard the anomaly as explained, and this article follows that view. But the 2012 result has stated limits, and it is more accurate to describe them than to oversell the finish.

The largest error source in the thermal model was the condition of the RTG surfaces: the performance of their coating after decades of exposure is unknown, and the paper assigns it a 25 percent uncertainty out of a 26.3 percent total. Heat escaping through instrument openings is poorly documented. The detailed 2012 analysis covered Pioneer 10 only, and Pioneer 11 was not treated in the same detail. The paper does not address an anomalous spin-down, and questions about when the effect begins (apparently near Saturn's orbit) and about periodic annual and diurnal terms in the data remain modeling issues, according to Wikipedia's summary.

Other teams reached related conclusions. Rievers and Lämmerzahl published high-precision thermal models in Annalen der Physik in 2011, and Bertolami and colleagues published related thermal work in 2012. Think of the situation as a court case: the thermal explanation does not require a new law of nature, it matches in direction and size, and it accounts for the observed decline. A rival theory would need to explain the same facts and more.

What the Pioneer anomaly taught us

The first lesson concerns precision. Tracking can detect forces near 10⁻¹⁰ m/s², about one eleven-billionth of Earth's surface gravity, which means solar-system tests of gravity can be demanding. Had a new force acted on the Pioneers at the level of the anomaly, it would be difficult to hide in planetary ephemerides, which was one reason the new-physics options kept losing ground. The same logic underlies precision tests such as the 1919 eclipse expedition and Einstein's theories of relativity, both of which had to survive increasingly exacting measurement.

The second lesson concerns engineering. A spacecraft is a thermal object as well as a mechanical one, and its heat has momentum. Missions designed for precise navigation now have to account for how their own heat leaves the vehicle, and the Pioneer case is the standard example. The same issue appears in missions that must survive hot and cold extremes, as discussed in how spacecraft survive extreme temperature swings.

The third lesson concerns how science handles surprises. The anomaly looked like a crack in the foundations of gravity, and it motivated a large literature on modified gravity. The resolution was an ordinary effect that nobody had computed with enough care. Both outcomes are possible, and it is hard to know which one you are in until the analysis is done. Compare 'Oumuamua, the interstellar visitor whose small non-gravitational acceleration also drew speculation, in our article on the 'Oumuamua mystery. Careful measurement is also what lets gravitational-wave astronomers trust signals at the edge of detectability, as described in gravitational waves explained, and what lets astronomers test gravity near extreme objects, as in black holes explained.

The takeaway

The Pioneer anomaly was real as a measurement and wrong as a hint of new physics. The probes were being pushed toward the Sun by about 8.74 × 10⁻¹⁰ m/s², and the push came from heat: waste heat from four plutonium generators and the electronics, radiated and reflected unevenly by the spacecraft's own shape. Solving it took rescued tapes, a hand-built thermal model with thousands of elements, and the patience to compare a model against telemetry and Doppler data. The finished answer is slightly untidy, with 80 percent agreement and open details, but it asks for no new force. The lesson is to take small residuals seriously and to check the boring explanations first, because the boring explanations are usually right and still take real work to prove.

Frequently Asked Questions

Q: What was the Pioneer anomaly? A: It was a small, steady acceleration of Pioneer 10 and Pioneer 11 toward the Sun, about (8.74 ± 1.33) × 10⁻¹⁰ m/s², detected as a residual in radio Doppler tracking at distances of roughly 20 to 70 AU. It appeared after known forces were subtracted, and it equals a speed change of roughly 1 km/h per decade.

Q: Who solved the Pioneer anomaly? A: A team led by Slava Turyshev, with Viktor Toth, Gary Kinsella, Siu-Chun Lee, Shing Lok, and Jordan Ellis, published the decisive analysis in Physical Review Letters in 2012. Earlier work by J. I. Katz, by Rievers and Lämmerzahl, and by Bertolami and colleagues supported the same thermal explanation from different angles.

Q: How can heat push a spacecraft? A: Light and thermal infrared radiation carry momentum, so a spacecraft that radiates more heat in one direction recoils the other way. For Pioneer, the force needed was about 2.3 × 10⁻⁷ newtons, which corresponds to roughly 68 watts of perfectly directed light, a few percent of the heat the plutonium generators produced at launch.

Q: Why did it take so long to solve? A: The anomaly was noticed around 1980 and not seriously investigated until 1994. Solving it then required recovering tracking and telemetry data from old tapes and building a thermal model of the spacecraft from blueprints, because no digital design files existed. The Planetary Society helped fund the data rescue, which was reported complete in 2006.

Q: Did the anomaly challenge general relativity? A: It was proposed as a possible hint of new gravitational physics, but the thermal explanation removes that need. Planetary ephemerides also constrained many gravitational alternatives, because a Sunward force of that size would show up in the orbits of the outer planets. The resolution is consistent with standard gravity.

Q: Why didn't the Voyager probes show the same anomaly? A: Voyager spacecraft rely on more frequent thruster firings for attitude control, which adds noise that makes tracking at the 10⁻¹⁰ m/s² level far harder. The Pioneers were spin-stabilized and used thrusters only occasionally, so their tracking was clean enough to expose the effect.

Sources and Further Reading

• Anderson, J. D. et al. (2002). "Study of the anomalous acceleration of Pioneer 10 and 11." Physical Review D, 65, 082004. arXiv abstract
• Turyshev, S. G., Toth, V. T., Kinsella, G., Lee, S.-C., Lok, S. M., and Ellis, J. (2012). "Support for the thermal origin of the Pioneer anomaly." Physical Review Letters, 108, 241101. arXiv:1204.2507 and full text
• Turyshev, S. G. and Toth, V. T. (2010). "The Pioneer anomaly." Living Reviews in Relativity, 13, 4. arXiv:1001.3686
• Katz, J. I. (1999). Comment on "Indication, from Pioneer 10/11, Galileo and Ulysses Data, of an Apparent Anomalous, Weak, Long-Range Acceleration." Physical Review Letters, 83, 1892. arXiv:gr-qc/9809070
• Markwardt, C. B. (2002). "Independent confirmation of the Pioneer 10 anomalous acceleration." arXiv:gr-qc/0208046
• Rievers, B. and Lämmerzahl, C. (2011). "High precision thermal modeling of complex systems with application to the flyby and Pioneer anomaly." Annalen der Physik, 523(6), 439.
• Milgrom, M. (1983). "A modification of the Newtonian dynamics as a possible alternative to the hidden mass hypothesis." Astrophysical Journal, 270, 365.
• NASA Science: Pioneer 10 mission page
• Astronomy magazine: How the Pioneer anomaly was solved
• Wikipedia: Pioneer anomaly, Pioneer 10, Modified Newtonian dynamics

---

Read next:

• Cosmic Rays and Particle Physics: Sources, Detection, and the Oh-My-God Particle
• Parker Solar Probe: Touching the Sun
• The Cassini Grand Finale: What 22 Death Dives Taught Us About Saturn