Mercury Explained: Iron Core, Polar Ice, Odd Orbit

Mercury is the smallest planet in the solar system and the one nearest the Sun, yet it keeps defying the expectations that location creates. Its core takes up roughly 83 to 85 percent of its radius. Its poles hold water ice despite daytime highs near 430 °C. And because it rotates exactly three times for every two trips around the Sun, one sunrise-to-sunrise day lasts 176 Earth days, which is twice as long as its 88-day year. NASA's MESSENGER probe mapped all this between 2011 and 2015, and ESA's BepiColombo is due to arrive in November 2026.

A planet that is mostly core

Mercury has a radius of about 2,440 km, a little over a third of Earth's width and only slightly larger than our Moon, according to NASA. Its mean density is 5.427 grams per cubic centimeter, nearly the same as Earth's 5.515, even though Mercury is far too small to squeeze its interior the way Earth's gravity does. That is the clue. A small planet with Earth-like density has to be made of much denser material, and the denser material is iron.

Models based on MESSENGER's measurements of the planet's spin and gravity put the metallic core at a radius of about 2,020 km, give or take 30. NASA's own page gives about 2,074 km, which it describes as roughly 85 percent of the radius. Either way, the core accounts for about 57 percent of Mercury's volume, against about 17 percent for Earth. A thin rocky layer, with mantle and crust together about 420 km thick, makes up the rest.

Property	Mercury	Earth (for contrast)
Radius	About 2,440 km	About 2.6 times larger
Mean density	5.427 g/cm³	5.515 g/cm³
Core share of volume	About 57%	About 17%
Core radius	About 2,020 to 2,074 km	Not comparable in proportion
Moons	None	One

Mercury is small and dense. Its density rivals Earth's despite far weaker gravity compression.

One way to picture it is a peach with a stone that has grown to nearly the full size of the fruit. The flesh is a thin layer. Or consider a hard-boiled egg where the yolk takes up most of the interior and the white is only a thin film.

How anyone knows what is inside a planet

Nobody has drilled into Mercury, so the core size comes from indirect measurements. The key quantity is the moment of inertia, a number that describes how mass is distributed from center to surface. A sphere with all its mass at the center has a very low value. A uniform sphere has a higher one. Margot and colleagues measured Mercury's spin in 2012 and derived a moment-of-inertia factor of 0.346 ± 0.014. Hauck and colleagues combined that with MESSENGER's gravity field in 2013 and estimated the top of the liquid core at a radius of 2,020 ± 30 km. A value that low says the mass is concentrated toward the center.

Radar observations of how Mercury wobbles as it turns also bear on the question. Research published in 2007 by a Cornell and National Radio Astronomy Observatory team suggested that at least part of the core is molten, and MESSENGER's later results were consistent with a liquid iron core. A liquid, moving core matters, because it is what a magnetic field needs. That is why Mercury has one at all, which comes up again below.

The comparison with Earth is instructive. Earth's inner layers were learned largely from earthquake waves, and Mercury has not had seismometers. The planet is being read the way a doctor reads a patient without surgery: from outside, through the way it responds to being pushed.

Why would a planet end up like this?

Nobody knows for certain. The question is one of the open puzzles of planetary science, and the main hypotheses were laid out in a 2014 Sky & Telescope summary by J. Kelly Beatty.

• Vaporization. Mercury started Earth-like, and a young, energetic Sun heated its outer layers until they largely vaporized.
• Metal-rich building blocks. Mercury formed from iron-rich material in the first place.
• A "big splat." Mercury was larger at birth, and a massive collision stripped off much of its crust and mantle.

A 2014 paper in Nature Geoscience by Erik Asphaug and Andreas Reufer offered a twist on the third idea. They proposed that Mercury was the smaller body in one or more glancing, hit-and-run blows, each of which stripped away some of its rock while leaving it intact. Asphaug described the outcome as "about one-in-10 lucky."

MESSENGER's chemistry changed the debate. The probe found notable amounts of sulfur, potassium, and sodium on the surface. These are volatile elements that boil away at high temperature, so a planet cooked by the young Sun should have lost them. In the Sky & Telescope account that finding doomed the vaporization idea and left the big splat "hanging by a thread." Later work has suggested that giant-impact scenarios are feasible but difficult, and the question remains open. A giant impact also appears elsewhere in solar system history: how the Moon was formed covers the best-supported case of a planet being reshaped by collision.

Broader context for how planets assemble from dust and gas, and why an iron-rich world could come out of that process, is in what a protoplanetary disk is.

MESSENGER: the first spacecraft to orbit Mercury

Before MESSENGER, Mercury had been seen up close only by Mariner 10, which flew past it in 1974 and 1975. A flyby gives a snapshot of part of a planet. An orbiter gives a mapping campaign.

MESSENGER, a NASA Discovery Program mission built by the Johns Hopkins Applied Physics Laboratory with Sean Solomon as principal investigator, launched on August 3, 2004. Getting a probe to Mercury is harder than it sounds. A spacecraft falling toward the Sun gains speed and has to shed it to be captured, so MESSENGER used gravity assists from one Earth flyby, two Venus flybys, and three Mercury flybys to slow down while saving propellant. It entered orbit on March 18, 2011, becoming the first spacecraft to orbit the planet.

The orbit was deliberately elongated. It brought the probe within about 200 km of the surface at the closest point and out to roughly 10,000 to 15,000 km at the far end (sources differ on the exact figure). Close to 100,000 images came back during the primary mission, which finished in March 2012, and the team had completed a global map by March 2013. After two extensions, MESSENGER ran out of propellant and was deliberately crashed into the surface on April 30, 2015.

Date	Event
1974 to 1975	Mariner 10 makes the first flybys
August 3, 2004	MESSENGER launches
2008 to 2009	Three Mercury flybys
March 18, 2011	MESSENGER enters orbit
March 2013	Global mapping completed
April 30, 2015	Planned impact ends the mission

MESSENGER's orbital phase lasted about four years.

Its major results were, in summary: water ice and possibly organic material in permanently shadowed craters near the north pole, water and other species in the exosphere, a magnetic field offset far to the north, evidence for a liquid iron core, signs of past volcanism, and fault scarps showing that the planet is still contracting. The sections below take those in turn. For a view of Mercury alongside the rest of the planets, see the planets of the solar system.

Why a planet this close to the Sun has ice

The idea sounds absurd. Mercury's daytime highs reach about 430 °C and its nights fall to about -180 °C, according to NASA. A planet that hot should be dry as a bone.

But there is a trick of geometry. Near Mercury's poles, some crater floors never see the Sun. Their rims block the light, and because the planet's rotation axis is almost upright relative to its orbit, the Sun never climbs high enough over the horizon to peer in. Those floors are permanently shadowed, and they stay extremely cold. Temperatures near the poles never exceed about 180 K, or -93 °C, and the shadowed floors themselves are colder still, below about 102 K, even though the rest of the planet bakes. It is the same principle as the cold, dark floors of craters near the Moon's poles.

The radar clue

In the early 1990s, Earth-based radar observations found bright patches near Mercury's poles. Radar reflects strongly off ice, so the patches suggested water ice in shadowed craters, though other explanations were possible. The radar-bright features all lie inside regions that stay shadowed, which fits the ice idea but did not prove it.

MESSENGER settled much of the question. Three papers published in Science (online in November 2012, in print in early 2013) reported that neutron measurements point to water ice as the major component of the north polar deposits. In the coldest places the ice is exposed at the surface. In others, a dark layer covers it, in places where it is slightly too warm for ice to be stable at the surface. David Lawrence, lead author of one paper, described a hydrogen-rich layer more than tens of centimeters thick beneath a surface layer of about 10 to 20 centimeters that is less rich in hydrogen. Paige's team attributed the dark material to complex organic compounds, possibly delivered by comets and volatile-rich asteroids. The delivery mechanism is discussed in asteroids, comets, and meteors.

How much ice?

An estimate puts the polar ice at roughly 10¹⁴ to 10¹⁵ kilograms, or 100 billion to 1 trillion tonnes across both poles. For scale, the Antarctic ice sheet holds roughly 2 to 3 × 10¹⁹ kilograms (my calculation from its volume of about 27 to 30 million cubic kilometers), so Mercury's total is tens of thousands of times smaller. It is a small reservoir on a planetary scale and an immense one on a human scale.

The honest summary is that the ice exists, is partly covered by dark material, and has an uncertain total mass.

The 3:2 lock that makes a day longer than a year

Mercury's most counterintuitive property is its relationship between spin and orbit. The sidereal rotation period, the time to turn once relative to the stars, is 58.646 Earth days. The orbital period is 87.969 days. Three times the first number is 175.9 days. Two times the second is 175.9 days. Mercury rotates exactly three times for every two orbits, a pattern called a 3:2 spin-orbit resonance.

The consequence is that the time from one sunrise to the next, the solar day, is about 176 Earth days. That is two Mercury years. A day on Mercury lasts longer than its year.

Radar observations by Gordon Pettengill and Rolf Dyce in 1965 showed that the rotation period was about 59 days. Giuseppe Colombo proposed that same year that the planet was locked in a 3:2 resonance rather than a 1:1 lock like the Moon's, and Mariner 10 data later supported that view.

Why 3:2 and not 1:1?

One factor is Mercury's strongly elliptical orbit, which has an eccentricity of 0.2056. Mercury swings from 46.0 million km from the Sun at perihelion to 69.8 million km at aphelion, which is a big swing. Tides raised by the Sun are therefore much stronger near perihelion than near aphelion, and that variation helps maintain the 3:2 resonance. A planet on a nearly circular orbit would be more likely to settle into a 1:1 lock.

A sunrise that stops and reverses

Spin and orbit speed do not stay matched through the whole orbit. The planet moves fastest near perihelion, and for a few days around that point its angular speed along the orbit exceeds its rotation rate. For an observer on the surface in the right place, the Sun appears to slow, stop, back up across the sky, and then resume its normal path. Parts of the planet therefore see a double sunrise or a double sunset. It is a good example of how a planet's calendar depends on more than its size. Two Earth-like spin rates would give a very different Mercury. Why Venus is hotter than Mercury covers how another feature of a planet, its atmosphere, matters at least as much as its distance from the Sun.

A shrinking planet

When Mercury's interior cools, it contracts. As it shrinks, the brittle crust has to accommodate the loss, and it does so by buckling and breaking into thrust faults, where one slab of crust is pushed up and over another. Mariner 10 and MESSENGER found these as long cliffs called scarps, some stretching up to 1,000 km and standing up to 3 km high. Estimates of the total shrinkage of the radius are in the range of about 1 to 7 km.

The surprise came from MESSENGER's final months in orbit. Its low-altitude images of the last 18 months revealed small thrust-fault scarps, less than 10 km long with only tens of meters of relief. Thomas Watters and colleagues reported in Nature Geoscience in 2016 that these scarps look fresh and conclude that Mercury is likely still contracting, so new faults may be forming now. Their small size, pristine appearance, and crosscutting of impact craters point to an age under 50 million years, which is recent in geological terms.

Compare that with Venus, where the crust is thought to renew itself through volcanism, as described in Venus's volcanoes. Mercury is a different case: a small planet whose body is still settling billions of years after it formed.

The pocked surface

The surface is covered with impact craters, and with no weather to erase them they accumulate. Researchers have identified 46 impact basins, including the Caloris Basin at about 1,550 km across. The general question of why airless bodies are so heavily cratered is addressed in why the Moon has so many craters.

The orbit that tested Newton

Mercury's orbit has a famous second chapter. Planets do not trace the same ellipse forever. Gravity from the other planets tugs on them, and the point of closest approach to the Sun, the perihelion, creeps slowly around the Sun. This is called perihelion precession.

In 1859, the French astronomer Urbain Le Verrier noted that Mercury's perihelion advanced faster than Newton's theory allowed, by about 38 arcseconds per century. Simon Newcomb revised the discrepancy to about 43 arcseconds in 1882. An arcsecond is 1/3,600 of a degree, so the excess is tiny. At 43 arcseconds per century, it would take roughly 3 million years to produce one extra full turn, by a back-of-envelope calculation.

Le Verrier proposed a hidden planet closer to the Sun than Mercury, which he called Vulcan. Astronomers searched for it for decades. W. W. Campbell's eclipse observations in 1908 effectively ended the search, and no evidence for Vulcan was ever found.

Einstein's answer

The solution came from a new theory of gravity. In 1915, Einstein's general relativity predicted an extra shift from the curvature of spacetime near the Sun, and it accounted for the missing amount quantitatively. Einstein wrote to Michele Besso that the perihelion motion was explained quantitatively and that Besso would be astonished.

The modern breakdown shows how the pieces fit. The observed total is 574.10 ± 0.65 arcseconds per century. The prediction adds up to 575.31, made of about 532.30 from the pull of the other planets, 42.98 from general relativity, 0.0286 from the Sun's slight flattening, and a very small negative term from a frame-dragging effect. Without relativity the total would fall about 43 arcseconds short, which is far outside the uncertainty.

Contribution to Mercury's perihelion shift	Arcseconds per century
Pull of other planets	532.30
General relativity	42.98
Solar oblateness	0.03
Lense-Thirring (frame dragging)	-0.002
Predicted total	575.31
Observed	574.10 ± 0.65

General relativity supplies the 43 arcseconds Newton's mechanics could not.

This was one of the first successes of general relativity, and it came before the 1919 eclipse expedition. How relativity shows up in everyday technology is covered in how GPS corrects for relativity.

A weak field with a strange offset

Mercury is small, and a small, once-hot planet might be expected to have lost its internal heat and dynamo long ago. But Mariner 10 found a magnetic field, and MESSENGER measured it carefully. The field is about 1 percent as strong as Earth's, around 300 nanotesla at the equator. Its tilt is about 10 degrees, similar to Earth's 11 degrees.

The strange feature is the offset. MESSENGER data place the magnetic equator roughly 480 km north of the geographic equator. Anderson and colleagues reported this offset in Science in 2011 (a value of 484 ± 11 km), and later analyses give a similar figure of about 479 km. Researchers describe the field as a dipole displaced northward by about 20 percent of the planet's radius.

The weak field still matters, because it gives the planet a magnetosphere that shapes how the solar wind reaches the surface. The exosphere is a thin layer of atoms rather than a true atmosphere, mostly oxygen, sodium, hydrogen, helium, and potassium, with a pressure of less than 0.5 nanopascals. Compare that with Mars's more substantial but still thin atmosphere in why Mars lost most of its atmosphere.

BepiColombo: the next orbiter

Two spacecraft now approach the planet together. BepiColombo, a joint mission of ESA and JAXA, launched on an Ariane 5 on October 20, 2018. It carries the Mercury Planetary Orbiter (MPO) from ESA and the Mercury Magnetospheric Orbiter, called Mio, from JAXA, both attached to a Mercury Transfer Module that provides propulsion for the trip.

The trip is long because the same problem that applied to MESSENGER applies here: slowing down. BepiColombo used flybys of Earth in April 2020, Venus in October 2020 and August 2021, and six flybys of Mercury itself, on October 1, 2021, June 23, 2022, June 19, 2023, September 4, 2024, December 1, 2024, and January 8, 2025.

Where things stand

According to ESA's arrival-phase updates, as of October 6, 2026, the transfer module separated on September 3, 2026, with all systems nominal. All MPO instruments were switched on together on September 15, and first analyses show them in good health. The BELA laser altimeter fired successfully and the SIMBIO-SYS camera took its first image in space. ESA lists no delays at this time. Orbit insertion is scheduled for November 21, 2026, the two orbiters are due to separate on December 9 and 10, 2026, and the science phase is to begin in April 2027.

The insertion had earlier been planned for December 2025. Reduced thruster power, traced to unexpected electric currents in the spacecraft's power system, pushed it to November 2026, according to Space.com's reporting. Dates for a mission still in flight can change, so anyone reading this after November 2026 should check ESA's pages.

What should BepiColombo add? A search with two orbiters can see the planet's field and its surroundings from two vantage points at once, which should help explain the offset and the way the magnetosphere responds to the solar wind. The orbiters' instruments are also expected to test the formation scenarios above by mapping surface composition more completely. More on how the mission fits into the current wave of exploration is in the most important space missions of the 2020s.

Things people get wrong about Mercury

"Mercury is the hottest planet." It is the closest, but Venus is hotter. Venus's thick carbon dioxide atmosphere traps heat, while Mercury's near-vacuum lets heat escape at night, so Mercury's temperature swings between about 430 °C and -180 °C.

"Mercury always shows the same face to the Sun." That describes a 1:1 lock, like the Moon's relation to Earth. Mercury is in a 3:2 resonance and rotates three times per two orbits, so every part of the surface eventually sees the Sun. That was the radar surprise in 1965.

"A planet that close to the Sun cannot hold water." It can, in the shadowed floors of polar craters, where temperatures stay far below the planet's daytime values. The ice is not an exception to the physics. It follows from it.

"Mercury is the biggest of the small worlds." Mercury is slightly larger than our Moon, but a couple of moons, such as Jupiter's Ganymede, are wider than Mercury. They are less massive, because Mercury is dense and iron-heavy. For the moon that outsizes it, see Ganymede, the largest moon.

"Mercury's orbit proves Einstein right on its own." The precession was an anomaly that general relativity explained, and modern tests rely on many more measurements. It was an important early success and not the only evidence.

FAQ

Q: Why does Mercury have such a large iron core? A: It is not settled. Hypotheses include a large impact that stripped its rocky outer layers, repeated glancing hit-and-run collisions, and formation from metal-rich material. MESSENGER's detection of volatile elements like sulfur and potassium on the surface ruled out simple vaporization by the young Sun, so the question remains open.

Q: Is there really ice on Mercury? A: Yes. MESSENGER found that water ice is the main component of deposits in permanently shadowed craters near the north pole, covered in many places by a dark layer that may be organic material. Reference estimates put the total at roughly 10¹⁴ to 10¹⁵ kilograms, far smaller than Earth's ice sheets.

Q: How long is a day on Mercury? A: One rotation relative to the stars takes about 58.6 Earth days, but a solar day, from one sunrise to the next, lasts about 176 Earth days. That is twice Mercury's 88-day year, because the planet rotates three times for every two orbits.

Q: Is Mercury shrinking? A: Evidence says so. As its interior cools, the planet contracts and the crust breaks into thrust-fault scarps. MESSENGER's last 18 months of images showed small, fresh scarps, and Watters and colleagues concluded in 2016 that Mercury is likely still contracting. Estimates of total radius loss range from about 1 to 7 km.

Q: What did Mercury have to do with Einstein? A: Mercury's perihelion advanced about 43 arcseconds per century more than Newtonian gravity could explain. Astronomers first suspected an undiscovered planet called Vulcan. In 1915, general relativity accounted for the extra shift without any new planet, which was an early confirmation of the theory.

Q: When does BepiColombo reach Mercury? A: ESA schedules orbit insertion for November 21, 2026, with the two orbiters separating on December 9 and 10, 2026, and science operations beginning in April 2027. ESA reported no delays as of its October 6, 2026 update, though future changes are possible.

The takeaway

Mercury looks like a boring cratered rock until you compare it to what its neighborhood predicts. It should be dry, and it holds ice. It should have lost its magnetic field, and it has one, offset toward the north. It should keep one face to the Sun, and it turns three times for every two orbits. Even its orbit carried a lesson, since a tiny discrepancy in its motion pointed toward a better theory of gravity. MESSENGER answered some of these puzzles and sharpened others, especially the origin of the huge core. BepiColombo, due to enter orbit in November 2026, is built to test those answers with two spacecraft working together. For the wider map of the solar system and beyond, the complete space guide is the place to continue.

Sources

• NASA, Mercury facts: size, day and year, temperatures, core, magnetic field, exosphere.
• Mercury (planet), Wikipedia: radius, density, core radius, 3:2 resonance, eccentricity, scarps, ice estimate, Caloris, magnetic field figures.
• MESSENGER, Wikipedia: launch, orbit insertion, imaging totals, key findings, end of mission.
• Sky & Telescope, Is Mercury a hit-and-run survivor? (July 10, 2014): formation hypotheses and MESSENGER's volatile chemistry.
• Tests of general relativity, Wikipedia: Le Verrier, Vulcan, Einstein, and the breakdown of the perihelion shift.
• ESA, BepiColombo factsheet and ESA, Latest updates: BepiColombo's arrival at Mercury: launch, flybys, separation, orbit insertion date, status as of October 6, 2026.
• Watters et al. 2016, Recent tectonic activity on Mercury revealed by small thrust fault scarps, Nature Geoscience 9, 743: small young scarps, age under 50 million years.
• JHU Gazette, Water ice found on Mercury: MESSENGER north polar ice papers in Science.
• Space.com, Thruster issues delay BepiColombo probe's Mercury arrival until November 2026: the delay from December 2025.