Planetary Science & Astrobiology

The big picture

Almost everything you are about to learn in this course was obtained without anyone touching the object being studied. Nobody has ever held a piece of Venus. No human has stood on Mars. The total mass of extraterrestrial material deliberately brought back to Earth by spacecraft amounts to a few hundred kilograms of Moon rock and a handful of grams from two asteroids and a comet's tail. And yet we can tell you the temperature at the base of Venus's clouds, the depth of the ocean under Europa's ice, and the fact that Mars once had rivers that flowed for thousands of years at a time.

That gap between how little we have touched and how much we claim to know is worth staring at before you accept a single further sentence of this course. Planetary science is an inferential discipline. Its conclusions are built out of photons, radio Doppler shifts, magnetic field measurements, a small pile of rocks, and a great deal of physics and chemistry done in laboratories on Earth. Some of those inferences are as solid as anything in science. Others are provisional, and a few famous ones have collapsed entirely. Learning to tell which is which is the single most useful skill this course can give you, and it matters most in the second half, where we get to astrobiology and the evidence gets thin and the stakes get emotional.

So we start with method. This lesson is about how planetary knowledge is actually produced: what a spacecraft can do, what light can tell you, what a rock in a laboratory can settle that no orbiter ever will, and where models fit in. If you have taken Introduction to Astronomy (ASTR 101) you already know how telescopes and spectra work in general; here we will focus specifically on how those tools are applied to solid, nearby bodies, which is a different game from studying stars.

Key idea: Planetary science is an inferential science built on remote measurement plus a very small physical sample, so every claim in it carries a specific and knowable strength of evidence, and your job as a student is to track that strength rather than just the headline.

The ladder of missions

Missions to other worlds come in a rough hierarchy. Each rung costs more, takes longer, and carries more risk than the one below it, and each rung answers questions the lower rungs cannot. Understanding this ladder explains why our knowledge of the solar system is so uneven: some worlds have been visited by everything up to a rover, and some have been photographed once, in passing, forty years ago.

Mission type	What it gives you	What it cannot do	Example
Flyby	First-order reconnaissance: images, a mass from gravitational deflection, basic spectra, magnetic field snapshot	See more than one hemisphere well, watch change over time, resolve seasons	Voyager 2 at Neptune, 1989
Orbiter	Global mapping, repeat coverage, gravity field, seasonal change, long baselines	Touch anything, measure below a few metres of surface directly	Magellan at Venus, 1990 to 1994
Atmospheric probe	Direct composition, pressure and temperature profile at one place and time	Represent the whole planet, survive long	Galileo probe into Jupiter, 1995
Lander	Ground truth at one point: soil chemistry, weather, seismology, imagery at human scale	Move to the interesting rock 40 metres away	Viking 1 and 2 on Mars, 1976
Rover	Choice of targets, stratigraphic context, drilling, in-place laboratory analysis	Match a terrestrial laboratory for precision or for asking new questions later	Curiosity and Perseverance on Mars
Sample return	Unlimited future analysis with the best instruments ever built, including ones not yet invented	Be cheap, be quick, avoid contamination worries	Apollo, Hayabusa2, OSIRIS-REx

Notice the pattern in that last column. Sample return is qualitatively different from everything above it because it removes the deadline. A rover's instruments were designed a decade before launch and cannot be upgraded. The Apollo lunar samples, by contrast, are still yielding new results in 2026 on instruments that did not exist in 1972, because the rocks are sitting in a curation facility in Houston waiting for the next technique. That single fact drives an enormous amount of mission planning, and it is why the stalled effort to bring the samples Perseverance has cached in Jezero Crater back to Earth is treated as such a serious loss by the community.

The ladder also explains a frustration you will meet repeatedly. We have four decades of continuous orbital coverage of Mars and exactly two brief flybys of Uranus and Neptune between them, both by the same spacecraft. That is not because ice giants are boring. It is because they are twenty years of flight time away and nobody has funded the trip.

Key idea: Every rung of the mission ladder buys a specific kind of knowledge, and the unevenness of what we know about the solar system tracks the unevenness of which rungs each world has received.

Remote sensing: what light will tell you

Most planetary data is light. The trick is that different wavelengths carry different information, and a well-designed instrument suite reads several at once.

Reflectance spectroscopy is the workhorse for surface composition. Sunlight hits a surface, most of it reflects, but specific wavelengths get absorbed by specific chemical bonds. Iron in a silicate mineral produces a broad absorption near 1 micrometre. Water bound in a mineral produces sharp features near 1.4 and 1.9 micrometres, and hydroxyl near 2.7. Carbonates, sulfates and clays each have their own fingerprints in the near infrared. So when the CRISM instrument on Mars Reconnaissance Orbiter found strong 1.9 and 2.2 micrometre absorptions in the layered rocks of Mawrth Vallis, that was not a guess about clay; it was a match to laboratory spectra of specific phyllosilicate minerals measured on a bench in Rhode Island. The inference chain is short and strong.

Thermal emission works the other way. Every warm object glows in the infrared, and the shape of that glow encodes temperature; departures from a smooth blackbody encode composition, since minerals emit differently at different wavelengths. Thermal data also gives you thermal inertia, which tells you whether a surface is fine dust, coarse sand, or bare rock, because those materials heat and cool at very different rates through a day. That is how landing site engineers know whether a spot is likely to swallow a lander.

Radar pierces things light cannot. Venus is permanently cloud-covered at visible wavelengths, so Magellan mapped 98 per cent of its surface by bouncing radar off it and timing the echoes, producing topography and roughness maps of a world nobody has ever seen with their eyes. Radar sounders can also see downward: the MARSIS and SHARAD instruments at Mars probe kilometres into the polar layered deposits.

Radio tracking may be the least glamorous and most quietly powerful technique in the field. You do not point it at anything. You simply measure the Doppler shift of the spacecraft's radio signal with extreme precision as it flies, and from the way its velocity changes you reconstruct the gravitational field it is moving through. That gives you the planet's mass, and then the lumps and bumps in that mass distribution, which is how we know Jupiter's core is fuzzy and how lunar mascons were discovered. Magnetometers and particle detectors round out the standard kit, and as you will see in Module 5, a magnetometer is how we first learned there is an ocean inside Europa.

Worked example: getting a planet's density from two numbers

Here is the simplest and most revealing calculation in planetary science, and it uses only quantities a flyby can deliver.

A spacecraft passing a planet gets deflected. Track the Doppler shift and you get GM, the gravitational parameter, to several decimal places; divide by the gravitational constant G and you have the mass. Image the disk, or watch the planet occult a star, and you get the radius. Two numbers, and now you can compute the bulk density, which is the first real constraint on what the world is made of.

Take Mars. Its mass is 6.417 x 10^23 kg and its mean radius is 3,390 km, which is 3.390 x 10^6 m.

Volume = (4/3) x pi x R^3 = (4/3) x 3.1416 x (3.390 x 10^6)^3

(3.390 x 10^6)^3 = 3.896 x 10^19 m^3, so Volume = 1.632 x 10^20 m^3.

Density = mass / volume = (6.417 x 10^23) / (1.632 x 10^20) = 3,932 kg/m^3, which is about 3.93 g/cm^3.

Now interpret it. Common silicate rocks have densities around 3.0 to 3.3 g/cm^3, and iron is about 7.9. A ball of pure rock the size of Mars would come in near 3.3 once you allow for compression at depth. Mars is denser than that, so it must contain a substantial amount of metal, but it is far from being mostly metal. Run the same arithmetic for Earth and you get 5.51 g/cm^3, which is much too high for rock alone even after compression, telling you immediately that Earth hides a big iron core. Run it for the Moon and you get 3.34, which is almost exactly the rock value and says the Moon has hardly any core at all. That one number, from two measurements, already sorts the inner solar system into categories, and it is the starting point for the interior modelling you will meet in Module 2.

Key idea: Mass and radius give bulk density, and bulk density alone divides the rocky worlds into iron-rich and iron-poor before a single spectrometer is switched on.

Rocks: sample return and the free samples

Everything above is inference at a distance. Physical samples are a different category of evidence, and there are two sources of them.

The first is sample return. Apollo and the Soviet Luna landers brought back about 382 kg of lunar material between 1969 and 1976, and China's Chang'e 5 and Chang'e 6 added more in 2020 and 2024, with Chang'e 6 returning the first samples ever collected from the Moon's far side. Hayabusa2 delivered 5.4 grams of the asteroid Ryugu in December 2020, and OSIRIS-REx delivered 121.6 grams of the asteroid Bennu in September 2023. Those are tiny masses that have produced an enormous scientific return, because a laboratory on Earth can do things no spacecraft instrument can: measure isotope ratios to parts per million, date minerals radiometrically, section a grain and image it at nanometre scale, and then repeat the whole thing differently when someone has a better idea in fifteen years.

The second source is free. Meteorites are pieces of other worlds that arrive without a launch vehicle, and there are more than 70,000 catalogued specimens, dominated by finds from Antarctica and hot deserts where dark rocks are easy to spot and slow to weather. Most come from asteroids. About 300 are recognised as lunar, and a similar number as Martian, identified because trapped gas in some of them matches the Martian atmosphere as measured by the Viking landers, which is a beautiful piece of cross-checking between two completely independent methods. The catch with meteorites is that you lose geological context: you have the rock but not the outcrop it came from, so you cannot say what was above it, below it, or next to it. Sample return buys context; meteorites buy quantity and variety for free.

Laboratories and models

Two more pillars complete the evidence structure. Laboratory analogue work reproduces planetary conditions on a bench: diamond anvil cells squeeze minerals to core pressures to see what phases form, cold chambers grow methane ice under Titan conditions, and shock guns fire projectiles into targets to study cratering. Every spectral identification mentioned above ultimately traces back to somebody measuring a known sample under controlled conditions and publishing the reference spectrum.

Numerical modelling ties it together, and it is where you should be most careful. Models of planet formation, atmospheric circulation, interior convection and impact physics are indispensable, since you cannot run an experiment on a planet. But a model is a hypothesis with arithmetic attached. It is only as good as its physics, its assumptions and its inputs, and models routinely produce confident-looking outputs that later turn out to have been wrong. The healthy relationship is that models generate predictions, observations test them, and disagreement is treated as information rather than as an inconvenience. When you read that a model shows Venus could have been habitable for two billion years, the correct response is to ask which assumptions did the work.

Key idea: The evidence chain runs from remote sensing to samples to laboratory calibration to models, and a claim is strong when several independent links point the same way, not when one link is loud.

A standard of evidence for the rest of the course

Because the second half of this course is about life, it is worth fixing three labels now and using them consistently.

Evidence is a measurement, reproducible and with stated uncertainty. Enceladus's plume contains molecular hydrogen at a few tenths of a per cent, measured by Cassini's mass spectrometer during a flyby in October 2015. That is evidence.

Plausible inference is a conclusion drawn from evidence using well-established physics or chemistry, where reasonable experts largely agree. That hydrogen most likely comes from water reacting with rock on the floor of Enceladus's ocean, because that reaction is well characterised on Earth and few alternatives fit. That is inference, and a strong one.

Speculation is anything that goes beyond what the inference supports. Since hydrothermal systems on Earth support thriving microbial communities, Enceladus might too. That is speculation. It is legitimate, it motivates missions, and it is not a finding. The problem in astrobiology is almost never that people speculate; it is that the three categories get flattened into one sentence by the time a result reaches the public. You will see exactly how that happens when we get to the Martian meteorite ALH84001 and to Venusian phosphine.

Common misconceptions

"We have basically explored the solar system." We have reconnoitred it. Every major body has been imaged at least once, which is a genuine achievement of the last sixty years, but reconnaissance is not exploration. Uranus and Neptune have each had one flyby. No spacecraft has ever landed on an ice giant moon, sampled Europa's ocean, or drilled more than a few centimetres into Mars.

"Pictures are the main scientific product." Images are the most shared product and they carry real science, especially geological context. But the results that change models most often come from spectrometers, magnetometers, mass spectrometers and radio tracking, none of which produce anything a newspaper would print.

"A rover can do what a laboratory does." A rover carries a few instruments chosen a decade earlier, operating on limited power, with no ability to prepare a sample the way a technician can. Curiosity's onboard laboratory is remarkable engineering and it is roughly comparable to a fraction of one bench in a university geochemistry department.

"If a model reproduces the observations, it is right." Several different models can reproduce the same observations, a situation known as non-uniqueness. It is common in gravity and interior modelling, where many internal density distributions fit the same external field. Models constrain; they rarely prove.

Recap

Planetary knowledge is produced by a ladder of missions, from flybys that give you a mass and a first look, through orbiters that map globally, to landers, rovers, and sample return, which removes the deadline on analysis. Remote sensing converts reflected light into mineralogy, thermal emission into temperature and grain size, radar into topography beneath clouds, and radio Doppler tracking into mass and interior structure. Mass and radius together give bulk density, and that single number already tells you that Earth has a large iron core, Mars a modest one, and the Moon almost none. Physical samples come from return missions and, for free, from meteorites, which trade geological context for quantity. Laboratory analogues calibrate every remote measurement, and numerical models tie the pieces together while remaining hypotheses with arithmetic attached. Throughout the rest of the course we will keep evidence, plausible inference, and speculation as three separate labels.

Sources

1. NASA Science. (n.d.). Solar system exploration. National Aeronautics and Space Administration. science.nasa.gov
2. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Overview of our planetary system. In Astronomy 2e (Section 7.1). OpenStax, Rice University. openstax.org
3. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Composition and structure of planets. In Astronomy 2e (Section 7.2). OpenStax, Rice University. openstax.org
4. NASA Astromaterials Research and Exploration Science. (n.d.). Astromaterials curation. NASA Johnson Space Center. curator.jsc.nasa.gov
5. USGS Astrogeology Science Center. (n.d.). Planetary geologic mapping. United States Geological Survey. astrogeology.usgs.gov
6. Encyclopaedia Britannica. (n.d.). Solar system. britannica.com

The big picture

Look at the solar system as a whole and the first thing you should notice is how orderly it is. All eight planets orbit in the same direction, the direction the Sun spins. All of them orbit in almost the same plane, within a few degrees of each other. The four inner planets are small, dense and rocky; the four outer ones are large, low-density and gas-rich. Nearly all the large moons orbit their planets in the same direction those planets rotate. And when you date material from all over the system, from lunar highlands to asteroid fragments to Martian meteorites, you get ages that cluster tightly around 4.5 billion years.

None of that is required by physics. You could imagine a system where planets orbited every which way at random inclinations, with rocky worlds and gas giants scattered without pattern, formed at wildly different times. That is not what we have. Order on this scale means common origin, and the job of a formation theory is to explain the order without hand-waving away the exceptions, of which there are also several important ones: Venus rotates backwards, Uranus is tipped on its side, and Earth has a Moon far too large for a planet its size.

This lesson tells the formation story as the field currently understands it, and flags the places where the story is genuinely unsettled. It matters for the rest of the course because almost every question we will ask later, about why Mars is dry, why Venus cooked, why Europa has an ocean, and what a typical planetary system looks like around other stars, is downstream of how planets get built.

Key idea: The solar system's shared plane, shared direction, compositional gradient, and shared age are the four observations that any formation theory has to produce, and together they point unavoidably to a single rotating disk.

From Kant and Laplace to a real disk

The core idea is old. Immanuel Kant proposed in 1755, and Pierre-Simon Laplace independently developed in 1796, that the Sun and planets condensed out of a single rotating cloud of gas. As the cloud contracted it would spin faster and flatten into a disk, and planets would form in that disk, which explains the shared plane and shared direction immediately. This is the nebular hypothesis, and in outline it survived.

In detail it nearly died. The problem was angular momentum. If the Sun formed from the centre of a collapsing rotating cloud, it should be spinning extremely fast. It is not. The Sun holds 99.8 per cent of the solar system's mass but only about 0.5 per cent of its angular momentum; Jupiter alone carries most of the rest. For a long stretch of the nineteenth and early twentieth centuries this discrepancy was serious enough that rival theories flourished, including one in which a passing star tore material out of the Sun. That rival is now dead, largely because it would make planetary systems vanishingly rare, and we now know they are common.

The modern resolution is that angular momentum gets transported outward through the disk by turbulence and magnetic coupling, and stripped from the young star by a magnetised stellar wind. The material that stays close spirals in and joins the star; the angular momentum ends up in the outer disk, and eventually in the giant planets. We observe the machinery directly: young stars are surrounded by disks that we can image, and they drive powerful outflows.

That last point deserves emphasis, because it converted the nebular hypothesis from a plausible story into an observed process. Protoplanetary disks around young stars in the Orion Nebula were imaged by Hubble in the 1990s. In 2014 the ALMA array in Chile returned an image of the disk around the young star HL Tauri showing a series of concentric dark gaps, exactly what you would expect if planets were already forming and clearing lanes, in a system only about a million years old. We are no longer inferring disks. We are photographing them.

The sequence, step by step

Step one: collapse. A dense core inside a cold molecular cloud, tens of thousands of astronomical units across and perhaps 10 K, becomes gravitationally unstable and begins to fall inward. It has some small initial rotation, and as it shrinks, conservation of angular momentum spins it up dramatically, the same effect as a skater pulling in their arms. Material falling along the rotation axis reaches the centre easily; material falling in the equatorial plane cannot, and piles up into a flattened protoplanetary disk.

Step two: heating and condensation. The centre becomes the protosun, and the inner disk gets hot, heated by the young star and by its own accretion. Everything volatile is vaporised near the star. As the disk cools, solids condense out in a strict order set by chemistry, called the condensation sequence. Refractory materials go first, at high temperature: aluminium and calcium oxides above about 1,500 K, then silicates and iron around 1,300 to 1,400 K, then sulfides and hydrated minerals lower down, and finally, below roughly 150 to 170 K, water ice. Further out still, ammonia and methane ices condense.

Step three: the frost line. Because temperature falls with distance from the star, the condensation sequence maps onto position. The frost line or snow line is the distance beyond which water ice is stable, and in the early solar system it sat somewhere around 2.5 to 3 astronomical units, out near the present asteroid belt. Inside it, the only solids available were rock and metal. Outside it, water ice was available too, and water is abundant. This is the single most consequential division in the solar system, because it roughly triples or quadruples the amount of solid material available to build planets beyond that line.

Step four: sticking things together. Micrometre dust grains collide gently and stick by electrostatic and surface forces, growing into millimetre and centimetre fluffy aggregates. Then there is trouble. At around a metre in size, objects both drift inward rapidly through gas drag, on timescales of centuries, and tend to shatter rather than stick when they collide. This is the metre-size barrier, and it was a genuine crisis in the theory for decades. The current favoured escape is the streaming instability: pebbles concentrate in dense clumps through their aerodynamic interaction with the gas, and when a clump becomes dense enough it collapses under its own gravity straight into a body tens or hundreds of kilometres across. Once you have such planetesimals, gravity takes over and growth accelerates.

Step five: from planetesimals to planets. Planetesimals gravitationally focus each other, and the biggest grow fastest in a phase called runaway growth, then settle into oligarchic growth where a few dozen Moon-to-Mars-sized embryos dominate their own zones. In the inner system, the final stage is violent: embryos collide with each other over tens of millions of years in giant impacts. The Moon-forming impact on Earth is the best-documented example we have of that final stage, and it is why Earth has an oversized satellite.

Key idea: Planet building runs dust to pebbles to planetesimals to embryos to planets, and the hardest step is the metre-size barrier, which the streaming instability is currently the best candidate for bypassing.

Worked example: where does water ice become stable?

You can locate the frost line with arithmetic you already have. For a body absorbing sunlight and re-radiating it, the equilibrium temperature at a distance a in astronomical units, for a perfectly absorbing object, is approximately

T = 278.6 / sqrt(a) kelvin.

Check it at Earth's distance: a = 1, so T = 278.6 K, about 5 degrees Celsius. That is the right ballpark for an airless Earth-like body with no greenhouse effect, and it is reassuring that a one-line formula lands there.

Now solve for where T falls to about 170 K, the temperature at which water ice becomes stable in a low-pressure nebula:

170 = 278.6 / sqrt(a), so sqrt(a) = 278.6 / 170 = 1.639, and a = 1.639^2 = 2.69 astronomical units.

That is essentially the middle of the asteroid belt, and it matches what the asteroid belt actually looks like: the inner belt is dominated by dry, stony S-type asteroids and the outer belt by dark, carbon-rich and water-bearing C-types. The belt preserves the frost line as a compositional boundary. Two cautions, though. The real frost line moved over time, since the disk cooled as it aged and as the young Sun's output changed, so it was farther out early and swept inward. And the formula ignores the disk's own optical thickness and internal heating, which shift the answer. Treat 2.7 AU as an order-of-magnitude anchor, not a surveyed boundary line.

Worked example: how much disk do you need?

Add up all the solid material in the planets and you get roughly 50 Earth masses, once you count the heavy-element cores of the giants along with the rocky planets. But the disk was made of Sun-like material, which by mass is about 71 per cent hydrogen, 27 per cent helium, and only about 1.4 per cent everything else. So to supply 50 Earth masses of solids you need at least

50 / 0.014 = 3,570 Earth masses of total nebular material.

The Sun is 333,000 Earth masses, so that is 3,570 / 333,000 = 0.0107 solar masses, about one per cent of a Sun. This is the classic minimum-mass solar nebula, and the word minimum is doing real work: accretion is inefficient and much of the disk is lost, so the true disk was heavier, probably several times this figure. The calculation is useful anyway, because it tells you the disk was a small fraction of the star, which is exactly what we observe around young stars today.

Making the giants, and moving them

Two competing routes to giant planets have been argued for decades. In core accretion, a solid core of perhaps 5 to 15 Earth masses assembles beyond the frost line, and once it is heavy enough its gravity captures hydrogen and helium from the disk faster and faster until the gas runs out. In disk instability, a massive cold disk fragments directly under its own gravity into gas clumps, skipping the core stage. Core accretion is the mainstream view for our giants, because it naturally explains why Jupiter and Saturn are enriched in heavy elements relative to the Sun, and because Juno's gravity measurements point to Jupiter having a heavy-element concentration at its centre. Disk instability remains a live possibility for some massive planets on wide orbits around other stars.

Core accretion has a timing problem you should know about. The core has to be built and the gas captured before the disk disperses, and observations of young star clusters show disks are gone by about 3 to 10 million years. Building a ten Earth mass core that fast by planetesimal accretion alone is difficult, which is a major reason pebble accretion, in which a growing core efficiently sweeps up centimetre-scale pebbles drifting inward through the gas, has become central to the modern picture.

Then there is migration, which is the biggest change to the story in fifty years and which the solar system alone would never have taught us. A planet embedded in a gas disk exchanges angular momentum with it and drifts, usually inward. Nobody took this seriously as a general phenomenon until 1995, when the first planet around a Sun-like star turned out to be a Jupiter-mass object orbiting 51 Pegasi every 4.2 days, far closer than Mercury. Such a planet cannot form where it is found, because it is inside the frost line and the disk there could not supply the material. It had to move. Within our own system, the Nice model proposes that Jupiter, Saturn, Uranus and Neptune formed closer together and later shifted, with Uranus and Neptune moving outward, scattering icy planetesimals and sculpting the Kuiper Belt. The related Grand Tack hypothesis has Jupiter migrating inward to about 1.5 AU and then back out, which would explain why Mars is so small and why the asteroid belt is so depleted. These models are constrained but not settled; treat them as leading hypotheses rather than established fact.

Key idea: Planets do not stay where they form, and the discovery of hot Jupiters around other stars forced migration from a curiosity into a standard ingredient of every formation model.

Common misconceptions

"The planets formed from material blown out of the Sun." They did not. Tidal and collision theories of that kind were taken seriously a century ago and are now abandoned. The planets and the Sun formed together from the same collapsing cloud, which is why the Sun's composition and the composition of primitive meteorites match so closely for non-volatile elements.

"Planets formed by gas slowly gathering onto a seed over billions of years." Formation was fast and violent. The gas disk lasted only a few million years, terrestrial planets finished assembling in tens of millions of years, and the final stage involved collisions between Mars-sized bodies, not gentle accumulation.

"The asteroid belt is a destroyed planet." It is not. The total mass of the belt is only about four per cent of the Moon, far too little to make a planet, and the objects in it have distinct compositions rather than looking like fragments of one differentiated body. The belt is material that never finished assembling, mostly because Jupiter's gravity kept stirring it to speeds where collisions shatter rather than stick.

"The frost line is a sharp line." It is a temperature condition that moved substantially as the disk evolved, and it applies to water specifically; other volatiles have their own condensation distances. Its position in the early solar system is inferred, not measured.

Recap

The solar system's shared orbital plane, shared direction, compositional gradient and common age of about 4.5 billion years all point to formation from a single rotating disk. Kant and Laplace had the idea; the angular momentum problem nearly sank it; the modern resolution moves angular momentum outward through the disk and out through stellar winds, and we now image protoplanetary disks directly, including gap-carving in HL Tauri. Solids condense in a temperature-ordered sequence, with the frost line near 2.7 AU dividing rock-only from rock-plus-ice territory, which you can verify with T = 278.6 divided by the square root of the distance in AU. Growth proceeds from dust to pebbles to planetesimals to embryos to planets, past the awkward metre-size barrier, probably via the streaming instability. Giants form by core accretion in the standard picture, racing a disk that lasts only a few million years, and pebble accretion helps them win that race. Finally, planets migrate, a fact the solar system hid from us and exoplanets revealed.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Origin of the solar system. In Astronomy 2e (Section 7.4). OpenStax, Rice University. openstax.org
2. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Formation of the solar system. In Astronomy 2e (Section 14.3). OpenStax, Rice University. openstax.org
3. NASA Science. (n.d.). Our solar system: Formation. National Aeronautics and Space Administration. science.nasa.gov
4. ALMA Partnership, Brogan, C. L., Perez, L. M., et al. (2015). The 2014 ALMA long baseline campaign: First results from high angular resolution observations toward the HL Tau region. The Astrophysical Journal Letters, 808(1), L3. doi.org
5. Mayor, M., & Queloz, D. (1995). A Jupiter-mass companion to a solar-type star. Nature, 378, 355-359. doi.org
6. Encyclopaedia Britannica. (n.d.). Solar system: Origin. britannica.com

The big picture

In the last lesson we said the solar system is about 4.5 billion years old and that formation was fast. Both of those are extraordinary claims about events that left no witnesses. This lesson is about how we know, and the answer is one of the genuine triumphs of twentieth century science: the rocks date themselves. Every meteorite that lands on Earth carries inside it a set of radioactive clocks that started running when its minerals crystallised, and if you can measure the isotopes precisely enough, the rock will tell you its own age to within a few hundred thousand years out of four and a half billion. That is a precision of better than one part in ten thousand, on an event that happened before there was an Earth.

The number that comes out is 4,567 million years, and it is not a rough estimate. The current best value for the oldest solids in the solar system is 4,567.30 million years with an uncertainty of about 0.16 million years, from lead isotope work on calcium and aluminium rich inclusions in the Efremovka meteorite published by Connelly and colleagues in 2012. Hold that in your head for a moment. We know when the solar system started forming, to within about 160,000 years, from a rock somebody picked up in Kazakhstan.

This lesson explains how that works, what the different kinds of meteorite sample, and what the resulting timeline actually says. It also handles a controversy, the so-called late heavy bombardment, where the evidence is genuinely more ambiguous than textbooks written twenty years ago suggest.

Key idea: Radiometric dating turns meteorites into self-recording clocks, and it is the only method that gives absolute rather than relative ages anywhere in planetary science.

A field guide to meteorites

Meteorites divide first into two philosophical categories, and the division matters more than the names.

Primitive meteorites never melted. Their parent bodies stayed small and cold enough that the material inside them was never sorted by density, so they preserve a mixture of dust, ice-derived minerals and condensates essentially as it was in the disk. These are the chondrites, named for the chondrules they contain: millimetre-sized once-molten silicate spheres that formed in seconds to minutes from flash heating events in the nebula, then cooled and were incorporated into their parent asteroid. Nobody is certain what did the flash heating. Shock waves in the disk and lightning-like electrical discharges are the leading candidates, and this is an honestly open question.

Chondrites come in families. Ordinary chondrites are the most common falls. Carbonaceous chondrites are the interesting ones for this course: dark, water-bearing, and rich in organic compounds including amino acids. The Murchison meteorite, which fell in Australia in 1969, has yielded more than 90 amino acids, most of which are not used by terrestrial life at all, which is exactly the pattern you would expect from abiotic chemistry rather than contamination. Enstatite chondrites are highly reduced and isotopically very close to Earth, which makes them important in arguments about what Earth was built from.

Embedded in some carbonaceous chondrites are calcium-aluminium-rich inclusions, or CAIs: white, irregular, refractory blobs a few millimetres across. These are the first solids that condensed as the disk cooled, and they are the oldest datable objects in the solar system. When you see the age 4,567 million years, it came from CAIs.

Differentiated meteorites are the opposite. Their parent bodies got hot enough to melt, and when a body melts, dense iron sinks and light silicate floats, exactly like slag and metal in a smelter. Break such a body up in a collision and you get achondrites from its rocky mantle and crust, iron meteorites from its metallic core, and stony-iron meteorites such as the beautiful pallasites from the boundary between the two. A slice of an iron meteorite, etched with acid, shows the Widmanstatten pattern, an interlocking crystal texture that only forms when nickel-iron cools at about one to a hundred degrees per million years. That texture is itself a measurement: it proves the metal cooled inside a body tens of kilometres across, insulated by rock.

Key idea: Chondrites sample material that never melted and therefore record the nebula itself, while achondrites and irons sample bodies that melted and separated, which means the solar system was building differentiated worlds almost immediately.

How radiometric dating actually works

A radioactive parent isotope decays into a stable daughter at a rate that is fixed, unaffected by temperature, pressure or chemistry at any conditions found in rocks. If you know the decay rate, and you can measure how much parent is left and how much daughter has accumulated, you can solve for the elapsed time.

The number of parent atoms remaining after time t is P = P0 x e^(-lambda x t), where lambda is the decay constant, related to the half-life by lambda = ln(2) / half-life. Every parent atom that decayed became a daughter atom, so D = P0 - P, and a little algebra gives the working equation:

t = (1 / lambda) x ln(1 + D / P)

That is the whole method in one line. Notice what it needs: the decay constant, which is measured in laboratories to high precision, and the present-day ratio of daughter to parent, which a mass spectrometer measures directly. It does not need to know the original amount of parent. That cancels out, which is why the method is robust.

Worked example: dating a meteorite with uranium and lead

Uranium-238 decays to lead-206 with a half-life of 4.468 billion years. First get the decay constant:

lambda = ln(2) / 4.468 = 0.6931 / 4.468 = 0.15513 per billion years.

Now suppose a mass spectrometer gives you a ratio of radiogenic lead-206 to remaining uranium-238 of exactly 1.000 in a mineral grain. Then

t = (1 / 0.15513) x ln(1 + 1.000) = 6.446 x 0.6931 = 4.468 billion years.

which is just a restatement that equal parent and daughter means exactly one half-life has passed. Good sanity check.

Now run it the other way to see what a CAI age requires. For t = 4.567 billion years:

D / P = e^(lambda x t) - 1 = e^(0.15513 x 4.567) - 1 = e^(0.7085) - 1 = 2.0309 - 1 = 1.031.

So a lead-206 to uranium-238 ratio of 1.031 corresponds to 4.567 billion years, while a ratio of 1.000 corresponds to 4.468 billion years. A three per cent difference in the measured ratio is a hundred million years of age. That sensitivity cuts both ways: it is why the method is powerful, and it is why the measurement has to be superb and why every real determination uses the two coupled uranium chains, uranium-238 to lead-206 and uranium-235 to lead-207, so that the two answers can be checked against each other. Disagreement between them, called discordance, flags a sample that has lost or gained lead and cannot be trusted.

Working geochronologists usually go further and use an isochron: measure several minerals from the same rock, plot daughter against parent normalised to a stable isotope of the same element, and fit a line. The slope gives the age and the intercept gives the initial daughter abundance, so you no longer have to assume the rock started with zero daughter. If the points do not fall on a line, the system was disturbed and you know not to trust it. Built-in failure detection is what separates a chronometer from a guess.

Short-lived clocks and a two-purpose isotope

Uranium and lead date the whole span. To resolve events separated by a million years, you need isotopes that die young. Aluminium-26 decays to magnesium-26 with a half-life of only 717,000 years, so after five million years less than one per cent remains, and after ten million it is essentially gone. Finding its decay products in a rock therefore proves the rock formed within a few million years of the aluminium-26 being made. This gives a fine-grained relative chronometer that ties CAIs, chondrules and early asteroid melting into a sequence.

Aluminium-26 has a second role that is arguably more important. Its decay released heat, and there was enough of it in the early solar system to melt any body larger than a few tens of kilometres across. That is the answer to a puzzle you might otherwise ask: how did asteroids, which are far too small to retain heat from accretion alone, get hot enough to form iron cores? They were cooked from the inside by a short-lived radioactive isotope, and the ones that formed a couple of million years later, after most of the aluminium-26 had decayed, never melted at all. The same isotope explains both why we have iron meteorites and why we also have chondrites.

Hafnium-182, decaying to tungsten-182 with a half-life of 8.9 million years, does a specific and elegant job. Hafnium prefers silicates and tungsten prefers metal, so when a body forms a core the two elements are separated, and the tungsten isotope signature of the mantle freezes in a record of when. Applied to iron meteorites, hafnium-tungsten dating says some asteroid cores formed within about one million years of CAIs. Applied to Earth, it says Earth's core formation was largely complete within roughly 30 to 100 million years.

The timeline that comes out

Event	Time after CAIs	Evidence
First solids (CAIs) condense	0, defined as 4,567.3 million years ago	Uranium-lead in CAIs
Earliest asteroid melting and core formation	Under about 1 million years	Hafnium-tungsten in iron meteorites
Chondrule formation	About 0 to 3 million years	Aluminium-26 and uranium-lead in chondrules
Gas disk disperses	About 3 to 10 million years	Disk lifetimes in young star clusters
Mars reaches roughly full size	A few million years	Hafnium-tungsten in Martian meteorites
Earth largely assembled, Moon-forming impact	About 30 to 100 million years	Hafnium-tungsten, lunar sample ages
Oldest surviving Earth minerals (Jack Hills zircons)	About 130 million years	Uranium-lead in zircon, roughly 4.4 billion years ago

Read that table as a whole and one thing should jump out: the interesting part was over fast. Everything from bare gas to a full set of planets took less time than the interval between the extinction of the dinosaurs and today. The remaining four and a half billion years were spent on cooling, cratering, and, on one planet, biology.

Key idea: The solar system built asteroids with molten iron cores within about a million years and finished the planets within about a hundred million, so planet formation is a brief, violent episode followed by a very long quiet period.

The late heavy bombardment: a contested chapter

Here is a case where you should watch a claim wobble. When Apollo samples were dated in the early 1970s, many impact melt rocks from different landing sites gave ages clustering near 3.9 billion years. Tera, Papanastassiou and Wasserburg proposed in 1974 that this reflected a terminal lunar cataclysm: a sudden spike in the impact rate roughly 600 million years after the Moon formed. When the Nice model appeared in the 2000s and showed that a late rearrangement of the giant planets could scatter a wave of planetesimals inward, the spike acquired a mechanism, and the late heavy bombardment became standard textbook material.

It has since been substantially challenged, on three grounds. First, sampling: all the Apollo and Luna sites lie on the near side within reach of ejecta from the Imbrium basin, so a cluster of 3.9 billion year ages may be measuring one enormous impact rather than a global spike. Second, resetting: a large impact can reset the isotopic clocks of older rocks, which biases the age distribution toward the most recent big event. Third, statistical reanalyses of lunar zircon and other datasets have found the apparent spike is not required by the data. Several groups now favour a declining bombardment, in which the impact rate fell steeply and monotonically from the end of accretion, possibly with a modest late enhancement, rather than a quiet period followed by a cataclysm.

What should you take away? That the Moon was heavily bombarded before about 3.8 billion years ago is not in doubt. Whether that bombardment was a discrete late spike is genuinely unsettled as of 2026, and settling it is one of the stated motivations for sampling far side terrain, which is exactly what China's Chang'e 6 mission did in 2024. This matters for astrobiology, because a cataclysm at 3.9 billion years would have sterilised or nearly sterilised Earth's surface, which would compress the window in which life arose. If the bombardment simply declined, that window is wider and more comfortable.

Common misconceptions

"Radiometric dating assumes a constant decay rate that might not be constant." Decay rates have been tested across enormous ranges of temperature, pressure, chemical state, and in astrophysical settings, and no variation relevant to geochronology has ever been found. More convincingly, different decay systems with wildly different half-lives and chemistries give the same ages for the same rocks, which they could not do if rates drifted.

"You have to know how much parent was there originally." You do not. The working equation depends only on the present daughter-to-parent ratio, and the isochron method also solves for the initial daughter abundance from the data.

"Carbon-14 is used to date meteorites." It is not. Carbon-14 has a half-life of 5,730 years and is useless beyond about 50,000 years. Deep-time dating uses systems such as uranium-lead, rubidium-strontium, samarium-neodymium, and potassium-argon, with half-lives in the billions of years.

"Meteorites are pieces of comets." Almost all recovered meteorites are asteroidal, with small numbers from the Moon and Mars. Cometary material is fragile and mostly burns up; the tiny particles that survive arrive as interplanetary dust rather than as hand samples, which is why the Stardust mission had to go and catch some.

Recap

Meteorites split into primitive chondrites, which never melted and preserve nebular material including chondrules and the refractory CAIs, and differentiated achondrites, stony-irons and irons from bodies that melted and separated a core. Radiometric dating uses t equal to one over the decay constant times the natural log of one plus the daughter-to-parent ratio, needs no assumption about starting amounts, and gains built-in error detection through isochrons and through comparing the two uranium-lead chains. CAIs date to 4,567.3 million years with an uncertainty near 0.16 million years. Short-lived aluminium-26 both fine-tunes the early chronology and supplies the heat that melted small bodies, while hafnium-tungsten dates core formation. The resulting timeline shows asteroid cores within about a million years and Earth largely assembled within about a hundred million. The late heavy bombardment, once standard, is now a contested interpretation of an Apollo sample set that may be dominated by ejecta from a single basin.

Sources

1. Connelly, J. N., Bizzarro, M., Krot, A. N., Nordlund, A., Wielandt, D., & Ivanova, M. A. (2012). The absolute chronology and thermal processing of solids in the solar protoplanetary disk. Science, 338(6107), 651-655. doi.org
2. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Meteorites: Stones from heaven. In Astronomy 2e (Section 14.2). OpenStax, Rice University. openstax.org
3. Boehnke, P., & Harrison, T. M. (2016). Illusory late heavy bombardments. Proceedings of the National Academy of Sciences, 113(39), 10802-10806. doi.org
4. NASA Science. (n.d.). Meteors and meteorites. National Aeronautics and Space Administration. science.nasa.gov
5. Encyclopaedia Britannica. (n.d.). Meteorite. britannica.com
6. United States Geological Survey. (n.d.). Radiometric dating and the geologic time scale. usgs.gov

The big picture

The deepest hole ever drilled into the Earth is the Kola Superdeep Borehole in Russia, which reached 12.26 kilometres before the rock got too hot and plastic to continue. The Earth's radius is 6,371 kilometres. So humanity's deepest penetration into its own planet is about 0.2 per cent of the way to the centre, which on a standard classroom globe would be a scratch you could not see. Everything you have ever been told about the Earth's core is inference.

And yet the inference is excellent. We know the outer core is liquid iron alloy, we know its radius to within a few kilometres, we know the inner core is solid, and we know roughly how fast it is growing. That confidence comes from having thousands of seismometers on a planet that produces earthquakes obligingly and often. For every other rocky body, we have less. The Moon has Apollo-era seismic data. Mars has data from a single seismometer that operated from 2018 to 2022. Venus and Mercury have none at all, and their interiors are reconstructed from gravity, rotation, and physics.

This lesson is about that reconstruction: why planets have layers in the first place, what measurements constrain those layers, why small worlds die young, and where magnetic fields come from. It is foundational, because interior state controls almost everything at the surface. Whether a planet has volcanoes, whether it has plate tectonics, whether it holds an atmosphere, and whether it has a magnetic field to shield that atmosphere all trace back to how much heat is inside and how it gets out.

Key idea: A planet's surface behaviour is downstream of its internal heat budget, so interior structure is the hidden variable behind volcanism, tectonics, magnetism, and even atmospheric survival.

Why planets have layers

A newly assembled planet is a chaotic heap of everything that was in its neighbourhood: metal, silicate, sulfides, and if it formed beyond the frost line, ice. If it stays cold, it stays a heap, and that is exactly what chondrite parent bodies did. But if it gets hot enough for the metal to melt, the heap stops being stable. Iron alloy is nearly twice as dense as silicate rock, so it sinks. Light material floats. The body sorts itself into a metallic core, a silicate mantle, and a thin low-density crust on top. This is differentiation, and it is irreversible.

Four things provide the heat.

• Accretional heating. Every impact during formation converts kinetic energy into heat. The bigger the body grows, the faster incoming material arrives, so heating accelerates as the planet gets larger.
• Core formation itself. Sinking iron releases gravitational potential energy. For Earth this alone was enough to raise the average temperature by roughly 2,000 kelvin, so differentiation, once started, feeds itself.
• Short-lived radioactivity. As we saw in the last lesson, aluminium-26 melted bodies only tens of kilometres across during the first few million years.
• Long-lived radioactivity. Uranium-238, uranium-235, thorium-232 and potassium-40 keep producing heat for billions of years, and they are the reason Earth is still geologically alive today.

Add tidal heating for bodies squeezed by a nearby giant planet, which we will save for the ocean worlds in Module 5, and you have the full list. Notice that the first three are all early. Only long-lived radioactivity and tides can sustain a planet over billions of years, which is why the question of whether a world is still active is really a question about its size and its orbit.

How we see inside without digging

Five independent handles constrain interior structure, and the strength of a conclusion depends on how many agree.

Bulk density, from the last module, gives the overall metal-to-rock ratio. It cannot tell you whether the metal is concentrated in a core or scattered through the body.

The moment of inertia factor can. Spin a sphere and how hard it is to spin depends on where its mass sits. Physicists express this as C divided by M R squared, where C is the polar moment of inertia. For a uniform sphere of constant density the value is exactly 0.4. Concentrate mass toward the centre and it drops below 0.4; the more centrally concentrated, the lower the number. You measure it from a body's response to torques, its precession, or its gravity field combined with rotation data.

Body	Bulk density (g/cm^3)	C/MR^2	What it implies
Uniform sphere (reference)	any	0.400	No concentration at all
Earth	5.51	0.3307	Large dense core, about 55 per cent of the radius
Mercury	5.43	0.346	Enormous core, roughly 83 per cent of the radius
Mars	3.93	0.364	Modest core, roughly half the radius
Moon	3.34	0.393	Very small core, only about 20 per cent of the radius

Look at Mercury in that table. Its bulk density is almost the same as Earth's despite being far smaller, which means it is not squeezed by self-gravity the way Earth is, which means its uncompressed density is much higher. Mercury is mostly iron. Explaining that is one of the outstanding problems in planetary science, and the leading candidates are a giant impact that stripped its rocky mantle, or vaporisation of outer silicate layers by the young Sun.

Seismology is the gold standard. Waves from quakes travel at speeds set by the material they pass through, and they reflect and refract at boundaries. Pressure waves cross liquid; shear waves do not, which is how the liquid outer core of Earth was discovered in 1906 from the shadow it casts. Apollo astronauts left seismometers on the Moon that ran until 1977, and NASA's InSight lander operated a seismometer on Mars from 2018 to 2022, detecting more than a thousand marsquakes. Those data give Mars a liquid core with a radius near 1,830 kilometres, a value later refined downward by analyses that argue for a molten silicate layer sitting on top of the core. That refinement is a good example of the field correcting itself with the same data and better modelling.

Gravity mapping, from spacecraft tracking, reveals internal mass anomalies. The GRAIL mission mapped the Moon's gravity field so precisely that it constrained crustal thickness to a few kilometres and showed the lunar crust is far more porous than anyone expected.

Tidal response is the fifth handle. A body that is partly molten flexes more under tidal forcing than a rigid one, and the amount of flex, expressed as the Love number, is measurable from orbit. This is how we know Mercury has a liquid outer core despite having no seismometer: the amplitude of its rotational libration is far larger than a solid body could produce.

Key idea: Bulk density gives composition, the moment of inertia factor gives concentration, seismology gives boundaries, gravity gives lateral variation, and tidal response detects liquid, so a well-studied interior rests on several independent legs.

Worked example: why small worlds die young

Heat is generated throughout a planet's volume and lost through its surface. So the fundamental ratio governing how long a body stays hot is surface area over volume. For a sphere,

A / V = (4 pi R^2) / ((4/3) pi R^3) = 3 / R.

The whole result is in that one line: the ratio scales as one over the radius, so smaller bodies have proportionally more surface for their volume and lose heat faster. Put numbers on it.

Earth, R = 6,371 km: A/V = 3 / 6,371 = 4.71 x 10^-4 per km.

Mars, R = 3,390 km: A/V = 3 / 3,390 = 8.85 x 10^-4 per km.

Moon, R = 1,737 km: A/V = 3 / 1,737 = 1.73 x 10^-3 per km.

So per unit of interior volume, Mars has 1.9 times as much cooling surface as Earth, and the Moon has 3.7 times as much. The Moon should therefore have run out of internal heat long before Mars, and Mars long before Earth, and that is exactly the observed order: lunar volcanism largely ended around 3 billion years ago, Martian volcanism dwindled to rare eruptions, and Earth is still erupting today.

Be careful not to overstate this. The ratio is a strong first-order rule, not a law, and it is not the whole story. Composition matters, because radioactive elements are not distributed identically. Heat transport mechanism matters enormously, since convection moves heat far faster than conduction and only starts when a layer is hot and deep enough. And tidal heating can override the rule entirely, which is why Io, smaller than the Moon, is the most volcanically active body in the solar system. Use 3/R to predict, then check whether something else is going on.

Getting the heat out

Interiors move heat two ways. Conduction passes energy along without moving material, and rock is a poor conductor, so conduction alone is hopeless for cooling a planet-sized object. Convection physically moves hot material upward and cold material down, and it is vastly more efficient. Earth's mantle is solid rock that nonetheless convects, creeping at centimetres per year over geological time, which is the engine behind plate tectonics.

The outermost shell of a rocky planet, the cold and brittle lithosphere, is defined by how heat moves rather than by chemistry: it is the layer that conducts because it is too cold and rigid to convect. Below it, the hotter asthenosphere deforms. A planet that has cooled a great deal has a thick lithosphere, which is why Mars can support Olympus Mons, a volcano 22 kilometres high that would sink into the mantle on Earth.

Dynamos and magnetic fields

A planetary magnetic field is not a fossil bar magnet. Iron loses permanent magnetism above its Curie temperature, around 1,043 K for pure iron, and planetary cores are far hotter than that. Fields are generated actively, right now, by a dynamo: moving electrically conducting fluid generating electric currents that sustain a magnetic field, which in turn organises the flow. Three conditions must all be met.

1. A layer of electrically conducting fluid, in rocky planets liquid iron alloy.
2. Convection in that layer, driven by heat escaping into the mantle or by chemical buoyancy as a solid inner core freezes out.
3. Rotation, to organise the flow into coherent helical patterns.

The comparative record is instructive. Earth has a strong field, about 25 to 65 microtesla at the surface, with all three conditions met. Mercury has a real but weak field, roughly one per cent of Earth's, discovered by Mariner 10 in 1974 and mapped by MESSENGER; it is genuinely dynamo-generated, which was a surprise for such a small, slowly rotating world. Venus has no detectable intrinsic field despite almost certainly having a liquid core, and the likely reason is condition two: without plate tectonics, Venus's mantle does not carry heat away from the core fast enough for core convection to run.

Mars is the most poignant case. It has no global field today, but in 1997 Mars Global Surveyor detected intense stripes of remanent magnetisation frozen into the ancient southern highlands, in places far stronger than anything in Earth's crust. That crust was magnetised by a field that no longer exists. Mars had a dynamo, and it shut down, probably before about 4.1 billion years ago. The consequences are the subject of Module 3, because losing your magnetic field changes how your atmosphere interacts with the solar wind. The Moon tells a similar story: no field now, but Apollo samples carry remanent magnetisation implying an ancient lunar dynamo, which is remarkable given how small the lunar core is.

Key idea: Magnetic fields require a conducting liquid, convection in it, and rotation, all at once, and Mars and the Moon both prove that a dynamo can start, run for hundreds of millions of years, and then stop for good.

Common misconceptions

"Earth's core is molten iron all the way through." The outer core is liquid; the inner core is solid iron, kept solid by pressure of around 330 gigapascals despite temperatures near 5,000 to 6,000 K. Pressure raises the melting point faster than the temperature rises with depth.

"A planet's magnetic field comes from magnetised rock." Only fossil crustal fields do, like those on Mars. Global fields are generated continuously by fluid motion, because core temperatures are far above the Curie point at which permanent magnetism is destroyed.

"Mars is dead because it is far from the Sun." Distance from the Sun has almost nothing to do with internal heat, which comes from accretion, differentiation and radioactivity. Mars cooled because it is small. Sunlight heats the surface, not the interior.

"Differentiation happened slowly over billions of years." For Earth, hafnium-tungsten dating puts core formation largely within the first 30 to 100 million years, and for small asteroids within about a million years. Differentiation is an early, fast process.

Recap

Rocky planets separate into core, mantle and crust when heat from accretion, core formation, and short-lived and long-lived radioactivity melts the metal so it can sink. We probe the result with bulk density for composition, the moment of inertia factor for how centrally concentrated the mass is, seismology for sharp boundaries, gravity mapping for lateral structure, and tidal response for detecting liquid layers. Mercury's moment of inertia factor of 0.346 combined with its high uncompressed density makes it an iron-dominated oddity. Because surface-to-volume scales as 3 over R, small worlds cool faster, which correctly orders the end of volcanism on the Moon, Mars and Earth, though tidal heating can override the rule. Convection, not conduction, does the real work of cooling a planet, and the rigid lithosphere is defined by where convection stops. Magnetic fields require a conducting liquid, convection, and rotation together, and both Mars and the Moon record dynamos that have since switched off.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Composition and structure of planets. In Astronomy 2e (Section 7.2). OpenStax, Rice University. openstax.org
2. Stahler, S. C., Khan, A., Banerdt, W. B., et al. (2021). Seismic detection of the Martian core. Science, 373(6553), 443-448. doi.org
3. NASA Jet Propulsion Laboratory. (n.d.). InSight mission. California Institute of Technology. jpl.nasa.gov
4. NASA Science. (n.d.). Mercury. National Aeronautics and Space Administration. science.nasa.gov
5. United States Geological Survey. (n.d.). Inside the Earth. usgs.gov

The big picture

Look at the Moon through even a cheap telescope and the dominant feature is holes. Look at Mercury, Callisto, Rhea, Phobos, Vesta or Ceres and you see the same thing. Impact cratering is the most widespread geological process in the solar system, the only one that operates on every solid surface regardless of size, temperature, composition or distance from the Sun. Volcanism needs internal heat. Erosion needs an atmosphere or a fluid. Tectonics needs a convecting interior. Cratering needs only that other objects exist, and they do.

For most of history, lunar craters were assumed to be volcanic. The impact interpretation was argued by Grove Karl Gilbert in 1893, resisted for decades, and only settled in the 1960s when Eugene Shoemaker demonstrated that Meteor Crater in Arizona contains shocked quartz phases that can only form at pressures far beyond anything volcanic. That is a nice illustration of the general point from Lesson 1: the argument was won in a laboratory, by a mineral, not by a photograph.

What makes cratering worth a whole lesson is that it is not just a destructive nuisance. It is a clock. Craters accumulate over time, so a surface with many craters is old and a surface with few is young. That single observation is the only way we have to date the surface of any world we have not sampled, which is to say almost all of them. This lesson teaches you to read that clock, work its arithmetic, and, just as importantly, know where it fails.

Key idea: Impact cratering is universal, and because craters accumulate with time, crater density is the only surface-dating tool available for worlds we have never sampled.

What actually happens in an impact

The first thing to understand is that an impact is not a collision in the everyday sense. Objects strike planets at typical speeds of 10 to 30 kilometres per second for the inner solar system, which is far faster than the speed of sound in rock, around 5 kilometres per second. The target has no time to get out of the way, and the projectile has no time to behave like a solid object. What happens instead is closer to an explosion.

Crater formation runs in three stages.

Contact and compression lasts a fraction of a second. Shock waves at pressures of hundreds of gigapascals propagate into both target and projectile. The projectile is almost entirely melted and vaporised. This is why almost nothing is ever found of the impactor at a large crater, a fact that confused nineteenth century investigators at Meteor Crater who went looking for a buried iron mass to mine and found nothing worth digging.

Excavation lasts seconds to minutes. The shock wave, now expanding roughly hemispherically, drives material outward and upward, opening a bowl-shaped transient cavity and hurling out an ejecta blanket. Fast-moving ejecta can travel hundreds or thousands of kilometres and produce bright rays, which are the streaks you see radiating from young lunar craters like Tycho.

Modification follows, as gravity pulls the oversteepened walls down. For small craters not much happens and you get a bowl: a simple crater, roughly a fifth as deep as it is wide, with a raised rim. Above a threshold diameter, the transient cavity is too big to hold itself up, the floor rebounds, and the walls collapse in terraces. The result is a complex crater: shallower, terraced, with a central peak. Larger still and the central peak becomes a ring of peaks, then multiple rings, producing the great impact basins.

The transition diameter between simple and complex depends on gravity, because gravity is what drives collapse. On the Moon it is about 15 to 20 kilometres. On Earth, with six times the surface gravity, it drops to about 3 kilometres. That relationship is a useful sanity check when you look at any new world: the transition tells you about the gravity, and it is another example of physics you can read off an image.

Worked example: how much energy is that?

Take a stony asteroid 1 kilometre across, so radius 500 metres, with a typical rocky density of 3,000 kg/m^3, hitting at 20 km/s.

Volume = (4/3) x pi x (500)^3 = (4/3) x 3.1416 x 1.25 x 10^8 = 5.24 x 10^8 m^3.

Mass = 3,000 x 5.24 x 10^8 = 1.57 x 10^12 kg, which is about 1.6 billion tonnes.

Kinetic energy = 0.5 x m x v^2 = 0.5 x (1.57 x 10^12) x (2.0 x 10^4)^2

= 0.5 x 1.57 x 10^12 x 4.0 x 10^8 = 3.14 x 10^20 joules.

To make that meaningful, one megatonne of TNT is 4.184 x 10^15 joules, so

3.14 x 10^20 / 4.184 x 10^15 = 75,000 megatonnes.

For comparison, the largest nuclear device ever detonated yielded about 50 megatonnes. A single kilometre-wide asteroid arrives with roughly 1,500 times that energy, and it would open a crater somewhere in the range of 10 to 20 kilometres across, since crater diameter typically runs 10 to 20 times the impactor diameter for these conditions. Notice how the arithmetic distributes the blame: mass scales as the cube of size, but energy scales as the square of speed, so speed matters enormously. Halving the size of that asteroid cuts the energy by a factor of eight; halving the speed cuts it by four. Both matter, and this is why a small fast object can outdo a larger slow one.

Key idea: Impacts release energy like explosions because kinetic energy goes as velocity squared at speeds far above the sound speed in rock, which is why the projectile vaporises and the crater is far larger than the impactor.

Reading the clock: crater counting

Here is the logic, in four steps, and it is worth being explicit because each step is an assumption you can check.

1. Impacts arrive at some rate per unit area per unit time.
2. A freshly resurfaced area, by lava or by a basin-forming impact, starts with zero craters.
3. Craters accumulate from that moment.
4. Therefore the number of craters per unit area is a measure of elapsed time since resurfacing.

Steps 2 through 4 are solid. Step 1 is where all the difficulty lives, because the rate has not been constant: it was much higher early on, when leftover planetesimals were still abundant, and it fell steeply. So the relationship between crater density and age is not a straight line but a curve, very steep before about 3.5 billion years ago and nearly flat since.

The curve has to be calibrated, and the only place it can be is the Moon, because the Moon is the only body from which we have samples with known locations and radiometric ages. Apollo and Luna gave us dated rocks from specific sites; we count craters at those same sites; and now we have pairs of crater density and absolute age that define the curve. Everything else in the solar system is dated by extrapolating from that lunar curve.

Astronomers usually quote crater density as N(1), the number of craters at least 1 kilometre in diameter per square kilometre. Here is a working table, rounded, built from the standard lunar chronology.

Surface age	N(1) per km^2	Craters at least 1 km wide per 1,000 km^2
3.9 billion years	3.3 x 10^-2	about 33
3.5 billion years	4.8 x 10^-3	about 4.8
3.0 billion years	2.6 x 10^-3	about 2.6
1.0 billion years	8.4 x 10^-4	about 0.84
0.5 billion years	4.2 x 10^-4	about 0.42

Worked example: date three surfaces

Suppose you are given orbital images and you carefully count every crater at least 1 kilometre across inside a mapped area of 5,000 square kilometres. Three regions give three results.

Region A: 165 craters. Density = 165 / 5,000 = 3.3 x 10^-2 per km^2. From the table, about 3.9 billion years. This is ancient highland-type terrain.

Region B: 24 craters. Density = 24 / 5,000 = 4.8 x 10^-3 per km^2. About 3.5 billion years. That is typical of the lunar maria, the dark basaltic plains.

Region C: 4 craters. Density = 4 / 5,000 = 8.0 x 10^-4 per km^2. Just under 1 billion years, so a young volcanic or resurfaced unit.

Now notice something important about Region A versus Region B. Region A has about seven times the crater density of Region B, yet it is only 400 million years older. That is the steepness of the early curve in action. In the ancient period, a small age difference produces a huge density difference, so ages are well separated. After about 3 billion years ago the curve flattens, and then the reverse is true: a large age difference produces only a small density difference, so young ages carry big error bars. A count that says a Martian lava flow is 200 million years old could easily mean 100 or 400 million. Take young crater-count ages as order-of-magnitude statements.

Also notice the statistics. Region C had four craters. The Poisson uncertainty on a count of four is roughly the square root, which is two, so the density is 8 x 10^-4 plus or minus 4 x 10^-4. That alone spans ages from about 500 million to 1.5 billion years. Counting more area is the only fix, and on small young units there may not be more area to count.

Key idea: Crater counting gives excellent relative ages everywhere and good absolute ages only where the lunar calibration applies, with uncertainties that grow badly for young surfaces and small counting areas.

Where the clock breaks

Saturation. Eventually a surface accumulates so many craters that each new one destroys an older one, and the count stops rising. The surface is then in saturation equilibrium, and it can tell you only that it is old, not how old. Much of the lunar highlands is at or near saturation at small crater diameters, which is why highland ages come from samples rather than counts.

Secondary craters. Ejecta thrown from a large primary impact lands hard enough to make craters of its own. These secondaries are numerous, small, often clustered or in chains, and they all formed at the same instant as the primary. Counting them as if they were independent primaries inflates the apparent age badly. Careful workers exclude obvious clusters, but distinguishing a scattered secondary from a small primary is genuinely difficult, and this remains a live source of disagreement about the ages of young Martian and lunar surfaces.

Extrapolation to other bodies. Mars sits next to the asteroid belt and therefore gets hit more often than the Moon, and by objects arriving at different speeds. Converting the lunar curve to a Martian one requires a modelled scaling factor, and the resulting absolute Martian ages carry an uncertainty of roughly a factor of two. Everyone in the field knows this; the numbers are still quoted, because they are the best available. When you read that a Martian outflow channel is 3.6 billion years old, mentally append a wide bracket.

Resurfacing that is not obvious. Atmospheres erode craters, ice flows, dust buries. On Venus, a thick atmosphere screens out small impactors entirely, so the smallest craters simply do not exist, and the counting must start at larger diameters. On Titan and Europa, active surfaces erase craters quickly, which is itself informative.

Common misconceptions

"The Moon has more craters because it gets hit more often." Per unit area the Moon and Earth are struck at similar rates. The difference is that Earth erases craters through plate tectonics, weathering, and burial, while the Moon preserves nearly everything. Earth has around 200 confirmed impact structures; the Moon has hundreds of thousands.

"Craters are mostly volcanic." This was a serious scientific position until the 1960s and it is now settled. Shocked mineral phases such as coesite and stishovite, plus shatter cones and impact melt, form at pressures no volcanic process reaches.

"A crater is about the size of the object that made it." Craters are typically 10 to 20 times the impactor diameter, because the energy released excavates far more material than the projectile itself contains.

"Crater counting gives precise dates." It gives robust relative ordering and calibrated absolute ages that are good on the Moon and factor-of-two elsewhere, with additional uncertainty from secondaries and small-number statistics.

Recap

Hypervelocity impacts behave like explosions, running through contact and compression, excavation, and modification, and producing simple bowl craters below a gravity-dependent transition diameter and terraced complex craters with central peaks above it. A 1 kilometre stony asteroid at 20 km/s delivers about 3.1 x 10^20 joules, roughly 75,000 megatonnes. Because craters accumulate on a surface that started clean, crater density measures age; the density-to-age curve is steep before 3.5 billion years and flat afterwards, and it is calibrated only by dated Apollo and Luna samples. Working through a table of N(1) values lets you convert counts directly into ages, with the caution that young surfaces and small areas carry large uncertainties. The method fails or degrades at saturation, in the presence of secondary craters, when extrapolated to other bodies, and where resurfacing has been subtle.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Dating planetary surfaces. In Astronomy 2e (Section 7.3). OpenStax, Rice University. openstax.org
2. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Impact craters. In Astronomy 2e (Section 9.3). OpenStax, Rice University. openstax.org
3. Neukum, G., Ivanov, B. A., & Hartmann, W. K. (2001). Cratering records in the inner solar system in relation to the lunar reference system. Space Science Reviews, 96, 55-86. doi.org
4. USGS Astrogeology Science Center. (n.d.). Planetary geologic mapping and crater analysis. United States Geological Survey. astrogeology.usgs.gov
5. NASA Science. (n.d.). Earth impact craters and near-Earth objects. National Aeronautics and Space Administration. science.nasa.gov

The big picture

Impacts come from outside. The two processes in this lesson come from inside, and they are what makes a world a place rather than a target. Volcanism moves molten rock from the interior to the surface. Tectonics deforms the surface by breaking, folding and moving it. Between them they build mountains, open basins, resurface continents, and vent the gases that make atmospheres. A world with no internal heat left has neither, and its surface is simply the accumulated record of impacts, which is roughly the situation on the Moon and Mercury today.

The comparison across worlds is where the science is. Every rocky body in the solar system has had volcanism at some point. Only one, as far as we know, has plate tectonics. Working out why that is true is not merely a curiosity about Earth. It bears directly on the astrobiology in the second half of this course, because plate tectonics runs the carbon cycle that has kept Earth's climate in a habitable band for billions of years, and if it turns out to be rare, that is a constraint on how common long-term habitable surfaces are.

Note before we start that this lesson covers Mercury, the Moon and Earth in detail and treats Venus and Mars comparatively. Those two get full lessons of their own in Module 3, because their atmospheres make them the two most instructive planets in the solar system for the questions this course cares about.

Key idea: Volcanism is ubiquitous across rocky worlds and plate tectonics appears to be unique to Earth, which makes the reason for that uniqueness one of the most consequential open questions in comparative planetology.

How rock melts, and what comes out

A planet's mantle is solid, but only just, and it takes surprisingly little to persuade some of it to melt. There are three ways.

Decompression melting is the most important. Rock's melting point rises with pressure, so hot mantle rising to shallower depths can cross its melting curve without gaining any heat at all. This is what happens under mid-ocean ridges and above rising mantle plumes, and it is the dominant melting mechanism in the solar system.

Flux melting works by adding water or other volatiles, which lowers the melting point of silicate rock substantially. On Earth this happens where a wet oceanic plate descends into the mantle, and it is why the volcanoes above subduction zones are so explosive. Without plate tectonics to carry water down, this mechanism is largely absent elsewhere.

Heat addition is the obvious one, and the least common, since it requires a genuinely new heat source such as tidal flexing.

What erupts depends mostly on silica content, which controls viscosity and therefore controls the shape of everything built.

Magma type	Silica	Viscosity	Typical landform	Where
Basaltic	Low, about 50 per cent	Runny	Broad shield volcanoes, flood basalt plains, lava tubes	Everywhere: Earth, Moon, Mars, Venus, Mercury, Io
Andesitic	Intermediate	Sticky	Steep stratovolcanoes, explosive eruptions	Earth subduction zones, essentially nowhere else
Rhyolitic	High, above 70 per cent	Very sticky	Domes, calderas, catastrophic ash eruptions	Earth continental crust

Read that table sideways and it tells you something important: basalt is the universal planetary lava. The dark lunar maria, the Martian volcanic plains, Venus's vast lava fields, and Mercury's smooth plains are all basaltic. The sticky, explosive, silica-rich magmas that produce Earth's most dramatic volcanoes require water-driven flux melting and repeated reprocessing of crust, which effectively requires plate tectonics. That is one concrete way Earth's tectonic style shows up in its rocks.

Tectonics: one planet does it differently

Every rocky body has a lithosphere. The question is whether that lithosphere is broken into pieces that move.

On Earth it is. The lithosphere is divided into about a dozen major plates that slide, collide, and sink, driven mainly by the weight of cold dense slabs pulling themselves down at subduction zones. New crust is created at mid-ocean ridges and destroyed at trenches, so the ocean floor is nowhere older than about 200 million years, a startling fact when the planet is 4.5 billion. Earth's surface is continuously recycled.

Everywhere else, so far as we can tell, the lithosphere is a single unbroken shell. This is the stagnant lid regime. The interior may still convect underneath, and volcanism can punch through, but the lid itself does not subduct. The consequences are visible from orbit. On a stagnant lid world a mantle plume stays under the same spot forever, so instead of a chain of volcanoes like Hawaii, where the plate moves over a fixed hotspot, you get one enormous volcano that keeps growing in place. Olympus Mons on Mars is 21.9 kilometres tall and about 600 kilometres across, roughly the footprint of Arizona, and its size is a direct signature of a lid that never moved.

Why does Earth alone break its lid? The honest answer is that we do not know for certain, and the leading hypotheses are water and size. Water weakens rock and lubricates faults, and Earth has plenty of it in its crust; Venus, which is nearly Earth's twin in size and density, is bone dry and has no plate tectonics, which is at least suggestive. Earth is also the largest rocky planet, which gives it the most internal heat to drive vigorous convection. Both explanations are plausible and neither is established, and with one example of plate tectonics in the known universe, we cannot do statistics.

Key idea: Stagnant lid worlds build single giant volcanoes because the crust never moves over the heat source, while plate tectonics recycles crust so thoroughly that no Earth ocean floor is older than about 200 million years.

Mercury: a shrinking world

Mercury looks superficially lunar in photographs, and that resemblance is misleading. It has extensive smooth plains, especially around and inside the Caloris basin, which MESSENGER data established are volcanic flood basalts rather than impact ejecta. It also shows evidence of explosive volcanism in the form of pyroclastic deposits, which is unexpected on a body assumed to be volatile-poor and remains an active puzzle.

Its signature feature, though, is tectonic and unlike anything else in the inner solar system: enormous lobate scarps, cliffs hundreds of kilometres long and up to about three kilometres high, that cut across craters and plains. These are thrust faults, where the crust has been pushed over itself. They are everywhere, and they all record the same thing. Mercury's huge iron core cooled and contracted, and the surface had to shrink to fit, wrinkling globally like the skin of a drying apple.

You can put a number on it. Mapping the scarps and adding up the shortening gives a radius decrease of roughly 7 kilometres, from an initial radius near 2,447 down to today's 2,440. The fractional change in surface area for a small change in radius is twice the fractional change in radius:

Delta A / A = 2 x (Delta R / R) = 2 x (7 / 2,440) = 0.0057, or about 0.57 per cent.

Just over half a per cent of surface area removed, and it produced cliffs three kilometres tall across an entire planet. That is a good lesson in how little global strain it takes to make dramatic landforms. Some of these scarps look crisp and cut small young craters, suggesting Mercury may still be contracting today.

The Moon: a two-terrain world

The Moon divides cleanly into bright, heavily cratered highlands and dark, smooth maria. The highlands are anorthosite, a rock made largely of plagioclase feldspar, and their existence is a major piece of evidence for the magma ocean hypothesis: the early Moon was molten to great depth, and as it crystallised, light plagioclase floated to form a global crust while denser minerals sank. The maria are flood basalts that flooded low-lying impact basins, mostly between about 3.9 and 3.0 billion years ago, though samples returned by China's Chang'e 5 mission in 2020 dated to about 2.0 billion years, extending known lunar volcanism considerably later than the Apollo record alone suggested.

Two more things are worth knowing. The near side and far side are strikingly different: maria cover about 31 per cent of the near side and only about 2 per cent of the far side, and the far side crust is thicker. The favoured explanation involves asymmetric crustal thickness concentrating heat-producing elements on the near side, and Chang'e 6, which returned the first far side samples in 2024, was aimed squarely at this problem. Second, the Moon is not tectonically dead: the Lunar Reconnaissance Orbiter has imaged small, crisp thrust faults called lobate scarps, geologically young, recording continued slow contraction, and shallow moonquakes recorded by Apollo seismometers may be associated with them.

Where did the Moon come from? The giant impact hypothesis holds that a Mars-sized body struck the proto-Earth about 4.5 billion years ago, and the debris coalesced into the Moon. It explains the Moon's tiny core, its depletion in volatiles, and the angular momentum of the Earth-Moon system. Its outstanding difficulty is isotopic: lunar and terrestrial oxygen, titanium and other isotopes are almost identical, which is hard to arrange if much of the Moon came from a different body. Variants involving a more energetic impact and thorough mixing are the current response, and the problem is not fully closed.

Earth as a planet

It is worth deliberately looking at Earth as one object in a comparative set rather than as home. Four features stand out, and only the first is unambiguously unique.

Plate tectonics, as discussed, appears to be Earth's alone. Abundant surface liquid water, covering 71 per cent of the surface, is unique among rocky bodies, though several icy moons have far more water than Earth does, just not on top. An oxygen-rich atmosphere at 21 per cent is unique and is a direct product of biology; oxygen is so chemically reactive that it cannot persist at that level without continuous resupply, which is precisely why it is a candidate biosignature in Module 6. And a large stabilising moon, which damps the wobble of Earth's axial tilt, is unusual for a rocky planet, though how much that matters for habitability is debated and often overstated.

The deepest connection between the geology in this lesson and the habitability questions later is the carbonate-silicate cycle. Carbon dioxide dissolves in rain, weathers silicate rock, and ends up as carbonate on the sea floor. Subduction carries that carbonate down; volcanism returns the carbon dioxide to the atmosphere. Crucially, weathering runs faster when the planet is warm and wet, so the cycle removes carbon dioxide faster when it is hot and slower when it is cold. That is a negative feedback, a thermostat operating on timescales of hundreds of thousands of years, and it is a leading explanation for how Earth stayed habitable while the Sun brightened by roughly 30 per cent over its lifetime. Notice that this thermostat needs both volcanism and subduction. Venus has the volcanism and not the subduction, and Module 3 will show you what that cost.

Key idea: The carbonate-silicate cycle links tectonics directly to climate stability, which is why plate tectonics is not just a geological curiosity but a candidate requirement for long-term habitability.

Common misconceptions

"Lava is molten rock from the Earth's core." It is not. The core is iron alloy and is far too deep. Nearly all magma originates in the upper mantle, within the top few hundred kilometres, mostly by decompression melting.

"Olympus Mons is huge because Mars has strong volcanism." Mars has less total volcanic output than Earth. Olympus Mons is huge because a stagnant lid kept the same patch of crust over the same plume for hundreds of millions of years, and because lower gravity and a thick cold lithosphere let a tall pile stand up.

"The Moon is geologically dead." Its volcanism ended long ago, but young thrust scarps and shallow moonquakes show it is still contracting and still faulting. Dead is too strong; quiet is right.

"Plate tectonics is the normal way planets work." It is the way exactly one known planet works. Every other rocky body we have examined operates in the stagnant lid regime.

Recap

Rock melts by decompression, by flux from added volatiles, or by heat addition, and silica content sets viscosity and therefore landform, with basalt the universal planetary lava and silica-rich explosive magmas essentially confined to Earth. Earth's lithosphere is broken into moving plates that recycle the entire ocean floor every 200 million years or so; every other rocky world has a stagnant lid, which is why Mars grew Olympus Mons to 21.9 kilometres rather than a moving chain of volcanoes. Mercury's global lobate scarps record about 7 kilometres of radius contraction, which is roughly a 0.57 per cent loss of surface area. The Moon divides into anorthositic highlands from a magma ocean and basaltic maria erupted mainly between 3.9 and 3.0 billion years ago, with Chang'e 5 pushing the record to 2.0 billion, and it still contracts today. Earth is distinguished by plate tectonics, surface water, biological oxygen, and a large moon, and its carbonate-silicate thermostat, which requires both volcanism and subduction, ties tectonics directly to long-term habitability.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). The lunar surface. In Astronomy 2e (Section 9.2). OpenStax, Rice University. openstax.org
2. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). The origin of the Moon. In Astronomy 2e (Section 9.4). OpenStax, Rice University. openstax.org
3. NASA Science. (n.d.). Earth's moon. National Aeronautics and Space Administration. science.nasa.gov
4. NASA Science. (n.d.). Mercury: Facts and exploration. National Aeronautics and Space Administration. science.nasa.gov
5. United States Geological Survey. (n.d.). Volcano hazards program: Types of volcanoes and lava. usgs.gov
6. Encyclopaedia Britannica. (n.d.). Plate tectonics. britannica.com

The big picture

An atmosphere is a shockingly thin thing. If you shrank the Earth to the size of a classroom globe, the entire breathable atmosphere would be a film thinner than a coat of varnish. More than half its mass sits below 5.5 kilometres, which is less than the height of many mountains. And yet that film sets the surface temperature, transports heat between equator and poles, shields the ground from ultraviolet light, allows liquid water to exist, and, on one world, contains the oxygen you are using to read this sentence.

Look at the inner solar system and the variation is extreme in a way that ought to be surprising. Venus, Earth and Mars are all rocky planets of broadly similar composition that formed in broadly the same part of the disk. Venus has a surface pressure of 92 bars, ninety-two times Earth's. Mars has 0.006 bars, about a two-hundredth of Earth's. Mercury and the Moon have essentially nothing. That is a range of more than four orders of magnitude between neighbours.

Explaining that range is the business of this lesson, and it comes down to three questions asked in order. Where did the gas come from? How much heat does it trap? And what makes it leave? Get those three straight and you can look at any world, in this solar system or another, and make a decent prediction about whether it holds air.

World	Surface pressure	Main gases	Mean surface temperature
Venus	92 bar	96.5 per cent carbon dioxide, 3.5 per cent nitrogen	737 K (464 C)
Earth	1.0 bar	78 per cent nitrogen, 21 per cent oxygen	288 K (15 C)
Mars	0.006 bar	95 per cent carbon dioxide, 2.8 per cent nitrogen	210 K (-63 C)
Titan	1.5 bar	94 per cent nitrogen, 5 per cent methane	94 K (-179 C)
Mercury and the Moon	Below 10^-14 bar	Traces of sodium, helium, hydrogen	Extreme swings

Key idea: Three neighbouring rocky planets span more than four orders of magnitude in surface pressure, so atmospheric outcome is not set by bulk composition but by the balance among supply, greenhouse trapping, and escape.

Where the air came from

There are three sources, and rocky planets mostly used the last two.

A primary atmosphere is hydrogen and helium captured directly from the nebula while the disk gas was still present. The giant planets kept theirs, which is why they are made of it. The rocky planets either never captured much or lost it almost immediately, because hydrogen is light, the inner disk was hot, and the young Sun's ultraviolet output was fierce.

A secondary atmosphere is outgassed from the interior. Volcanic eruptions release water vapour, carbon dioxide, sulfur dioxide, and nitrogen that were chemically bound in the mantle. This is the main source for Venus, Earth and Mars, and it explains why atmospheric history is tied to volcanic history, and therefore to planet size, all the way back to the 3 over R rule from Module 2.

Delivered volatiles arrive from outside, on comets and water-bearing asteroids. How much of Earth's water was delivered rather than outgassed is a real and unfinished argument. The evidence turns on deuterium-to-hydrogen ratios, a kind of isotopic fingerprint: Earth's ocean water has a D/H ratio close to that measured in some carbonaceous chondrites and unlike most measured comets, including comet 67P studied by Rosetta, which came in about three times Earth's value. The current lean is that asteroidal material supplied most of Earth's water, with comets a minority contributor, but this is an active area rather than a closed one.

Worked example: the greenhouse effect, measured three times

Here is the most useful calculation in comparative planetology. Start by computing what a planet's temperature would be with no atmosphere at all, purely from the sunlight it absorbs and the heat it radiates. The result is the equilibrium temperature:

T_eq = 278.6 x (1 - A)^0.25 / sqrt(a)

where A is the Bond albedo, the fraction of incoming sunlight reflected straight back, and a is the distance from the Sun in astronomical units. Then compare with the measured surface temperature. The difference is the greenhouse effect, measured rather than modelled.

Earth. A = 0.31, a = 1.0.

(1 - 0.31)^0.25 = 0.69^0.25 = 0.911. T_eq = 278.6 x 0.911 / 1 = 254 K.

Measured surface temperature is 288 K, so Earth's greenhouse effect is 288 - 254 = 34 K. Without it, the mean surface would be about -19 degrees Celsius and the oceans would freeze.

Venus. A = 0.77, a = 0.723.

(1 - 0.77)^0.25 = 0.23^0.25 = 0.693. sqrt(0.723) = 0.850. T_eq = 278.6 x 0.693 / 0.850 = 227 K.

Stop and look at that. Venus's equilibrium temperature is 227 K, which is colder than Earth's 254 K, because Venus's brilliant white clouds reflect 77 per cent of the sunlight that reaches them. Yet its measured surface temperature is 737 K. The greenhouse effect on Venus is 737 - 227 = 510 kelvin.

Mars. A = 0.25, a = 1.524.

(0.75)^0.25 = 0.931. sqrt(1.524) = 1.234. T_eq = 278.6 x 0.931 / 1.234 = 210 K.

Measured mean surface temperature is about 210 K, so Mars's greenhouse effect is only around 5 K. Its atmosphere is 95 per cent carbon dioxide, the same gas that cooks Venus, but there is so little of it that it barely warms the planet at all.

The lesson in those three numbers is that greenhouse warming depends on the amount and vertical structure of absorbing gas, not simply on its presence. Same gas, three planets, warming of 510 K, 34 K, and 5 K.

Key idea: Venus receives less absorbed sunlight per square metre than Earth does, yet is 483 kelvin hotter at the surface, which proves the difference is trapping rather than proximity to the Sun.

Worked example: escape velocity

Whether a planet keeps its air starts with how hard it is to leave. The escape velocity from the surface of a body of mass M and radius R is

v_esc = sqrt(2 G M / R), with G = 6.674 x 10^-11 in SI units.

Earth: M = 5.972 x 10^24 kg, R = 6.371 x 10^6 m.

2 G M = 2 x 6.674 x 10^-11 x 5.972 x 10^24 = 7.972 x 10^14.

Divide by R: 7.972 x 10^14 / 6.371 x 10^6 = 1.251 x 10^8. Square root: 11,190 m/s, so 11.19 km/s.

Mars: M = 6.417 x 10^23 kg, R = 3.390 x 10^6 m. 2 G M = 8.566 x 10^13; divided by R gives 2.527 x 10^7; square root is 5,027 m/s, so 5.03 km/s.

Titan: M = 1.345 x 10^23 kg, R = 2.575 x 10^6 m. This gives 2.64 km/s.

Titan's escape velocity is half of Mars's, and yet Titan has an atmosphere 250 times thicker at the surface. So escape velocity alone cannot be the answer, and that is the point of the next calculation.

Worked example: which gases actually escape

Gas molecules have a spread of speeds set by temperature, and the light ones move faster. The most probable speed of a molecule of mass m at temperature T is

v_th = sqrt(2 k T / m), with k = 1.381 x 10^-23 J/K.

Only the fastest molecules in the tail of the distribution exceed escape velocity, so the working rule of thumb used by planetary scientists is that a planet retains a gas over the age of the solar system if

v_esc is greater than about 6 x v_th

evaluated at the exobase, the altitude where the atmosphere becomes so thin that a molecule moving upward will not hit anything else. That is the right place to evaluate it, and it is usually much hotter than the surface because it absorbs ultraviolet light directly.

Earth, exobase near 1,000 K. For molecular hydrogen, m = 3.35 x 10^-27 kg:

2 k T / m = (2 x 1.381 x 10^-23 x 1000) / 3.35 x 10^-27 = 8.24 x 10^6, so v_th = 2,871 m/s = 2.87 km/s.

Six times that is 17.2 km/s, which exceeds Earth's 11.19 km/s. Earth loses hydrogen, and it does.

For molecular nitrogen, m = 4.65 x 10^-26 kg: 2 k T / m = 5.94 x 10^5, so v_th = 771 m/s. Six times that is 4.6 km/s, comfortably below 11.19. Earth keeps nitrogen, and it does.

Titan, exobase near 175 K. For nitrogen: 2 k T / m = 1.04 x 10^5, v_th = 322 m/s, six times is 1.93 km/s, below Titan's 2.64 km/s. Titan keeps nitrogen despite feeble gravity, purely because it is so cold. Cold is a substitute for mass.

Mars, exobase near 250 K. For atomic hydrogen, v_th is about 2.03 km/s and six times that is 12.2 km/s, far above Mars's 5.03, so Mars loses hydrogen fast. But for atomic oxygen, m = 2.66 x 10^-26 kg: 2 k T / m = 2.60 x 10^5, v_th = 509 m/s, six times is 3.06 km/s, which is below 5.03 km/s. On this criterion Mars should retain oxygen and carbon dioxide thermally.

And that is the punchline. Mars plainly did lose most of its atmosphere, but ordinary thermal escape cannot account for it. Something else had to do the work.

The other ways to lose an atmosphere

Thermal escape is only the first entry on a longer list.

Photochemical escape is the main culprit at Mars. Ultraviolet light ionises molecules high in the atmosphere; when an ionised carbon dioxide or oxygen molecule recombines with an electron, the reaction releases energy that can fling the resulting atoms away at above escape speed. This does not care about the average temperature at all.

Sputtering happens when energetic ions from the solar wind slam into the upper atmosphere and knock neutral atoms out like a cue ball breaking a rack.

Ion pickup occurs when an atom is ionised in a region where the solar wind's magnetic field passes through, and the newly charged particle is simply carried away by that field.

Hydrodynamic escape, sometimes called blowoff, is the violent case: intense extreme-ultraviolet heating drives a bulk outflow, a planetary wind, in which the escaping light gas drags heavier species along with it. This is thought to have been important for young planets around young active stars, and it is directly observed today around some hot exoplanets.

Impact erosion blasts atmosphere off entirely during large collisions, and was probably significant during the heavy bombardment.

Now the question everyone asks: does a planet need a magnetic field to keep its atmosphere? The popular version says Mars lost its air because its dynamo shut down, and that story is partly right and frequently overstated. A magnetosphere does deflect the solar wind and reduce sputtering and ion pickup, and MAVEN has measured Martian escape directly and watched it spike during solar storms. But Venus has no intrinsic magnetic field at all and has the thickest atmosphere of any rocky planet in the solar system. Venus is also more massive, has a higher escape velocity, and generates an induced magnetosphere from the interaction of the solar wind with its ionosphere. The honest summary is that a magnetic field is one factor among several, with gravity, temperature, and resupply from volcanism all mattering at least as much, and the field's protective role, while real, is smaller than most popular accounts suggest.

Key idea: Atmospheric retention is a budget problem set by gravity, upper-atmosphere temperature, the specific gases present, and continued volcanic resupply, with a magnetic field acting as a partial shield rather than as a switch.

Common misconceptions

"Venus is hot because it is closer to the Sun." Its equilibrium temperature, 227 K, is colder than Earth's 254 K, because its clouds reflect 77 per cent of the incoming light. The entire 510 kelvin excess is greenhouse trapping.

"The greenhouse effect is inherently bad." Earth's natural 34 kelvin of greenhouse warming is what keeps the oceans liquid. What is under discussion in climate science is a rapid human-driven increase on top of that, not the effect itself.

"Mars lost its atmosphere because it has no magnetic field." That is a contributing factor, not a complete explanation. Mars's low escape velocity of 5.03 km/s, its small size and consequent early loss of volcanic resupply, and photochemical escape all matter, and Venus's thick atmosphere without any intrinsic field is the standing counterexample.

"Escape velocity determines whether a gas escapes." It sets the bar, but what matters is the comparison between escape velocity and the thermal speed of the specific molecule at the exobase temperature. Titan, with half Mars's escape velocity, keeps a far thicker atmosphere because it is far colder.

Recap

Rocky planets built their atmospheres mainly by outgassing, supplemented by delivered volatiles whose deuterium-to-hydrogen fingerprint currently favours asteroidal over cometary sources for most of Earth's water. Equilibrium temperature, computed as 278.6 times the fourth root of one minus the albedo divided by the square root of the distance in AU, gives 227 K for Venus, 254 K for Earth and 210 K for Mars, and comparing those with measured surface temperatures yields greenhouse effects of 510, 34 and 5 kelvin respectively. Escape velocity, the square root of 2GM over R, is 11.19 km/s for Earth, 5.03 for Mars and 2.64 for Titan. Comparing escape velocity with six times the thermal speed at the exobase correctly predicts that Earth loses hydrogen and keeps nitrogen, that cold Titan keeps nitrogen despite weak gravity, and that Mars should thermally retain oxygen, which forces us to invoke photochemical escape, sputtering, ion pickup, hydrodynamic escape and impact erosion. A magnetic field helps, but Venus proves it is not decisive.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). The massive atmosphere of Venus. In Astronomy 2e (Section 10.3). OpenStax, Rice University. openstax.org
2. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Divergent planetary evolution. In Astronomy 2e (Section 10.6). OpenStax, Rice University. openstax.org
3. NASA Science. (n.d.). Planetary atmospheres. National Aeronautics and Space Administration. science.nasa.gov
4. NASA Goddard Space Flight Center. (n.d.). Planetary fact sheets. National Space Science Data Center. nssdc.gsfc.nasa.gov
5. Altwegg, K., Balsiger, H., Bar-Nun, A., et al. (2015). 67P/Churyumov-Gerasimenko, a Jupiter family comet with a high D/H ratio. Science, 347(6220), 1261952. doi.org

The big picture

If you were handed a list of planetary properties with the names removed, you would pair Venus with Earth immediately. Venus has 95 per cent of Earth's radius, 81.5 per cent of its mass, a bulk density of 5.24 against Earth's 5.51, and it formed 28 per cent closer to the Sun out of the same part of the same disk. On the crude measures available to an astronomer looking at an alien system, Venus and Earth are the same planet twice.

The surface of Venus is 737 kelvin, hot enough to melt lead. The atmospheric pressure is 92 bars, equivalent to being 900 metres underwater. The clouds are droplets of concentrated sulfuric acid. The longest any spacecraft has survived down there is 127 minutes, achieved by the Soviet Venera 13 lander in March 1982, which returned colour images of a flat, orange, rock-strewn plain before it died.

That is the most important comparison in planetary science, and it is why Venus gets a lesson of its own in a course about the conditions for life. Two nearly identical planets, next door to each other, one of which has hosted a biosphere for at least three and a half billion years and one of which is a furnace. If you want to know how fragile habitability is, this is the experiment nature already ran.

Key idea: Venus is the closest thing we have to a controlled experiment on habitability, because it holds size, composition, and formation location roughly fixed and varies mainly the solar flux and the history of water.

The numbers that set up the problem

Start with how much sunlight Venus gets. The solar constant at Earth is 1,361 watts per square metre, and flux falls as the inverse square of distance, so at Venus's 0.723 AU:

S = 1,361 / (0.723)^2 = 1,361 / 0.5227 = 2,604 watts per square metre.

That is 1.91 times what Earth receives, which sounds decisive. But now compute what each planet actually absorbs, which means multiplying by one minus the albedo and dividing by four to average over the whole sphere.

Venus: (1 - 0.77) x 2,604 / 4 = 0.23 x 651 = 150 watts per square metre.

Earth: (1 - 0.31) x 1,361 / 4 = 0.69 x 340 = 235 watts per square metre.

Venus absorbs only 64 per cent as much energy as Earth does. Its clouds are so reflective that, on the energy books, it is the cooler planet. Every one of the 483 kelvin by which it exceeds Earth's surface temperature has to be produced by trapping, not by supply. Hold onto that, because it is the single fact that most people get wrong about Venus.

A few other numbers matter. Venus rotates retrograde, backwards relative to its orbit, once every 243 Earth days, which is longer than its 225 day year. It has no moon and no intrinsic magnetic field. Its atmosphere, however, rotates far faster than the planet does: the cloud tops circle Venus in about four days, a phenomenon called superrotation that is still not fully explained and that JAXA's Akatsuki orbiter, in operation since 2015, was built to study.

The runaway greenhouse

Here is the mechanism, and it is a feedback loop rather than a single event.

Suppose you start with a young Venus that has liquid water, as most models of volatile delivery say it should have had. Water vapour is itself a powerful greenhouse gas. Warm the planet slightly and more water evaporates; more water vapour traps more heat; that warms the planet further, which evaporates more water. This is a positive feedback, and unlike the negative feedback of the carbonate-silicate cycle, it does not settle down. Past a threshold in absorbed flux, the loop runs away: the oceans evaporate entirely into the atmosphere.

That alone would give you a wet, hot planet. What makes the loss permanent is the second stage. With no cold trap to stop it, water vapour reaches the upper atmosphere, where solar ultraviolet light breaks it into hydrogen and oxygen. Hydrogen is the lightest gas there is, and it escapes readily, as your calculation in the last lesson showed even for Earth. The oxygen left behind is chemically greedy and reacts with iron in surface rocks and with volcanic gases. Both components of the water are removed, and the process is irreversible.

Is there evidence, or is this just a plausible story? There is evidence, and it is isotopic. Deuterium is heavy hydrogen, and being twice as massive it escapes less efficiently than ordinary hydrogen, so a planet that has lost hydrogen to space becomes enriched in deuterium. Pioneer Venus measurements in 1978 and later work found that Venus's deuterium-to-hydrogen ratio is roughly 150 times Earth's. That is a very large enrichment, and the natural interpretation is that Venus lost an enormous amount of hydrogen, meaning an enormous amount of water.

Be careful about the next step, though, because this is exactly where evidence turns into inference. Converting a D/H enrichment into an original water inventory requires a model of how the escape fractionated the two isotopes over billions of years, and different assumptions give answers ranging from a global ocean a few metres deep to one hundreds of metres deep. The claim that Venus was once wet is strong. The claim that it had an Earth-sized ocean is model-dependent, and you should treat any specific depth you see quoted with suspicion unless the assumptions come with it.

Key idea: The deuterium enrichment of about 150 times terrestrial is solid evidence that Venus lost a great deal of hydrogen, and therefore water, but the size of the original ocean is a model-dependent inference rather than a measurement.

Why Venus and not Earth

The extra sunlight started it, but the reason Venus could not recover is the failure of the thermostat you met in Module 2. Earth's carbonate-silicate cycle works like this: carbon dioxide dissolves in rainwater, the resulting weak acid weathers silicate rock, the products wash to the sea and are deposited as carbonate, and subduction eventually returns that carbon to the interior for volcanoes to re-emit. Weathering runs faster when it is warm and wet, so the cycle removes carbon dioxide faster when the planet is hot. It is a thermostat, and it needs liquid water to operate.

Remove the water and the thermostat stops. Volcanoes keep supplying carbon dioxide and nothing takes it out, so the gas accumulates in the atmosphere for billions of years. That is why the Venusian atmosphere is 96.5 per cent carbon dioxide at 92 bars.

Now for the comparison that makes the whole thing click. Earth has a comparable total amount of carbon dioxide to Venus. It is simply not in the air. It is locked in limestone and other carbonate rocks in the crust, roughly enough to produce a Venus-like atmosphere if it were all released at once. The difference between the two planets is not how much carbon they have; it is where the carbon is kept. Earth has water, so it files its carbon in rock. Venus lost its water, so its carbon sits in the sky.

Was Venus ever habitable?

This deserves a straight answer: nobody knows, and the question is genuinely open as of 2026.

One line of work, most prominently a 2016 modelling study led by Michael Way at NASA Goddard, ran three-dimensional climate models of an early Venus with a shallow ocean and its slow rotation, and found that the slow rotation produces thick dayside cloud cover that shades the planet enough to keep surface temperatures moderate. In those models Venus could have had liquid surface water for perhaps two to three billion years, only losing it relatively recently in solar system history.

A competing line, notably work led by Martin Turbet published in 2021, argues that the young Venus never got the chance. In their models the atmosphere stays too hot for water vapour to condense in the first place, so Venus emerges from its magma ocean phase already steamy and never forms an ocean at all.

Both are climate models. They disagree because they make different assumptions about cloud behaviour, which is the hardest thing in atmospheric science to get right even for Earth. Neither is evidence in the sense we defined in Lesson 1, and settling the argument requires measurements Venus has not yet received: the noble gas and isotopic composition of the deep atmosphere, and the mineralogy of the oldest terrain. Both are on the target list for the next generation of missions, which is a good example of a scientific dispute directly generating a mission requirement.

The surface, and how young it is

Magellan mapped 98 per cent of Venus by radar between 1990 and 1994, and the picture is of a volcanic world. About 80 per cent of the surface is smooth volcanic plains. There are broad shield volcanoes, strange flat-topped pancake domes formed from viscous lava under high pressure, and thousands of coronae, circular fracture systems tens to hundreds of kilometres across, thought to form where mantle upwellings push against the crust. The oldest and most deformed terrain, ridged and intensely fractured, is called tessera.

The crater record is the strange part. Venus has fewer than a thousand impact craters, and they are distributed almost randomly, with very few obviously degraded ones. On a planet with a long, steady geological history you would expect a range of crater ages and states of preservation. A near-random distribution of fresh craters implies the whole surface is roughly the same age, somewhere between about 300 million and 1 billion years, which for a planet is startlingly young.

Two interpretations compete. The catastrophic resurfacing hypothesis says that Venus, lacking plate tectonics to bleed heat steadily, builds up internal heat until it overturns in a planet-wide volcanic episode, resets the surface, and goes quiet again. The equilibrium resurfacing hypothesis says that ongoing regional volcanism, spread over time, can produce a similar statistical signature without any global catastrophe. The debate is not settled, and it matters, because a planet that periodically resurfaces itself is a different kind of object from one that renews itself steadily.

Is Venus active now? The evidence has strengthened. In 2023 Robert Herrick and Scott Hensley reported in Science that a volcanic vent in Magellan radar images had visibly changed shape and area between observations taken in February and October 1991, the first direct detection of ongoing volcanic activity on Venus. Venus Express infrared data had earlier suggested some lava flows are geologically very young. The current view is that Venus is very probably volcanically active today, which is a substantial change from the textbook position of twenty years ago.

Key idea: Venus's near-random population of fewer than a thousand craters means its entire surface is only 300 million to 1 billion years old, and direct detection of a changing volcanic vent in Magellan data indicates the planet is still active.

Exploring a place that destroys spacecraft

Venus has been visited more than most people realise, mainly by the Soviet Union. Venera 7 made the first successful landing on another planet in 1970, transmitting for 23 minutes. Venera 9 returned the first image from the surface of another planet in 1975. Venera 13 and 14 in 1982 returned colour panoramas and analysed soil. The American Pioneer Venus mission arrived in 1978, Magellan mapped the planet in the early 1990s, ESA's Venus Express operated from 2006 to 2014, and JAXA's Akatsuki has studied the atmosphere since 2015.

Then, for nearly three decades, almost nothing. Three missions are currently in development: NASA's DAVINCI, an atmospheric descent probe, NASA's VERITAS, a radar and infrared mapper, and ESA's EnVision. All three have been subject to schedule and budget pressure, and as of early 2026 their launches are planned for the early 2030s rather than the late 2020s originally advertised. That is a snapshot, and it is exactly the kind of number that moves; check a current mission page rather than trusting a date in any course, including this one.

Case study: the phosphine claim

In September 2020, a team led by Jane Greaves published in Nature Astronomy a reported detection of phosphine, PH3, in the Venusian cloud deck at an abundance of about 20 parts per billion. The reason it made headlines is that on Earth phosphine in an oxidising environment is produced almost exclusively by biological activity or by industry, and the authors stated they could not identify an abiotic route that would generate the observed amount on Venus. Cloud-level Venus, at around 50 kilometres altitude, has temperatures and pressures close to Earth-surface values, so it is the one region of the planet where life has ever been seriously proposed.

Watch what happened next, because this is how the process is supposed to work.

First, the data were re-examined. The detection rested on a single spectral line fitted against a strongly curved instrumental baseline using a high-order polynomial, and several groups showed that such a fit can manufacture apparent features from noise. Second, the observatory found a calibration error in the ALMA dataset; reprocessing reduced the inferred abundance substantially, and the original team's revised figure came down to around 1 part per billion, in the low single digits rather than 20. Third, a separate concern was raised that a sulfur dioxide line sits very close in frequency to the phosphine line, and Venus has abundant sulfur dioxide, so the identification itself was questioned. Fourth, independent attempts produced mixed results: some observations reported nothing, a 2022 report using the SOFIA airborne observatory described a tentative detection, and later ground-based campaigns have reported variable results.

Meanwhile the interpretation was attacked from the chemistry side. Venus's clouds are concentrated sulfuric acid, which destroys phosphine very quickly, so any phosphine present would need vigorous continuous resupply. Other groups proposed abiotic sources including volcanic delivery of phosphides. And even setting all of that aside, phosphine on Venus would be an anomaly requiring explanation, not a detection of life.

So what does the field conclude in 2026? That the presence of phosphine on Venus is unconfirmed. It is neither established nor definitively refuted, and it is certainly not evidence of life. Sort it into the three labels from Lesson 1: the original measurement was evidence of a spectral feature; that the feature is phosphine is a contested inference; that phosphine implies life is speculation. The episode was not a failure of science. Claims were published, tested hard by people with no stake in the answer, and revised. It also had a real effect: it revived scientific and public interest in Venus and strengthened the case for the missions now in development.

Common misconceptions

"Venus is hot because it is closer to the Sun." Venus absorbs only about 64 per cent as much energy per square metre as Earth. The heat is entirely a trapping effect.

"Venus proves what carbon dioxide will do to Earth." The physics of greenhouse trapping is the same, but a genuine runaway requires evaporating an entire ocean, which needs vastly more absorbed flux than Earth receives or will receive from human emissions. Venus is a useful illustration of feedbacks and a poor analogy for near-term terrestrial climate change.

"Phosphine was found on Venus, so there may be life there." The detection is unconfirmed after calibration corrections, possible line confusion with sulfur dioxide, and mixed independent results. Even a confirmed detection would be an unexplained anomaly rather than a biosignature.

"Venus is geologically dead." The young surface age, infrared evidence for fresh lava, and the 2023 detection of a vent that changed shape in Magellan imagery all point to a planet that is still volcanically active.

Recap

Venus is nearly Earth's twin in size and density but absorbs less energy per square metre because its clouds reflect 77 per cent of incoming sunlight, so its 737 kelvin surface is produced entirely by a 510 kelvin greenhouse effect. A runaway greenhouse evaporated its water, ultraviolet light split the vapour, and hydrogen escaped, leaving a deuterium enrichment about 150 times terrestrial as the isotopic fingerprint. Without liquid water the carbonate-silicate thermostat failed, so volcanic carbon dioxide accumulated in the air rather than being filed away in rock as it is on Earth. Whether Venus was ever habitable is genuinely unresolved, with competing climate models disagreeing about early cloud behaviour. Its surface is 300 million to 1 billion years old with fewer than a thousand craters, and it is very probably still volcanically active. The 2020 phosphine report is a clean case study in how a biosignature claim gets tested, and its current status is unconfirmed.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). The geology of Venus. In Astronomy 2e (Section 10.2). OpenStax, Rice University. openstax.org
2. NASA Science. (n.d.). Venus. National Aeronautics and Space Administration. science.nasa.gov
3. Greaves, J. S., Richards, A. M. S., Bains, W., et al. (2021). Phosphine gas in the cloud decks of Venus. Nature Astronomy, 5, 655-664. doi.org
4. Herrick, R. R., & Hensley, S. (2023). Surface changes observed on a Venusian volcano during the Magellan mission. Science, 379(6638), 1205-1208. doi.org
5. European Space Agency. (n.d.). EnVision. esa.int
6. Encyclopaedia Britannica. (n.d.). Venus. britannica.com

The big picture

Mars is the most thoroughly explored world besides Earth. More than twenty spacecraft have operated there, we have had continuous orbital coverage for over two decades, six rovers have driven across the surface, and we have a global topographic map more accurate than the map of Earth's ocean floor.

The reason for that attention is a single, well-supported conclusion: Mars used to be wet. Not damp, not occasionally frosty, but genuinely wet, with rivers that ran and lakes that stood for long periods. That conclusion is now about as secure as anything in planetary science, built from independent lines of evidence that agree. And it creates the question this lesson is really about. Earth had liquid water and life appeared within a few hundred million years. Mars had liquid water at the same time, next door, made of similar material. Did anything happen there?

We do not know, and that has been the honest answer for fifty years. What has changed is the quality of the question. This lesson walks the evidence ladder for water, explains where the atmosphere went, and then works through the search for life, including the most instructive false alarm in the history of astrobiology.

Key idea: That Mars once had abundant liquid surface water is firmly established by several independent lines of evidence, and everything about the search for life there follows from that single fact.

The planet in numbers, and its three ages

Mars has 53 per cent of Earth's radius, 10.7 per cent of its mass, and an escape velocity of 5.03 km/s. Its atmosphere is 95 per cent carbon dioxide at about 6 millibars, roughly a two-hundredth of Earth's, which is below the triple point of water over most of the planet: liquid water exposed at the surface today would boil and freeze at the same time. Mean surface temperature is about 210 K.

Martian history divides into three epochs, defined by crater counting.

Epoch	Approximate age	Character	Diagnostic minerals
Noachian	Before about 3.7 billion years ago	Heavy bombardment, valley networks, lakes, probable warm and wet intervals	Clays (phyllosilicates), formed in neutral water
Hesperian	About 3.7 to 3.0 billion years ago	Massive volcanism, catastrophic outflow channels, increasingly acidic water	Sulfates, formed in acidic brines
Amazonian	About 3.0 billion years ago to now	Cold, dry, wind-dominated, occasional volcanism	Anhydrous iron oxides, the rust that makes Mars red

Read that table as a story and it says the climate deteriorated in stages. This mineral sequence, mapped largely by the OMEGA and CRISM orbital spectrometers, is one of the strongest arguments that the change was gradual rather than a single event.

The evidence ladder for water

Rung one: shapes. Mariner 9 in 1971 revealed branching valley networks in the ancient highlands that look like terrestrial river systems, and enormous outflow channels up to 100 kilometres wide, carved by catastrophic floods. Orbiters later found deltas, layered fan-shaped deposits where valleys enter craters, which form only where a river runs into standing water. Jezero Crater has a textbook example, and that is why Perseverance was sent there. But morphology alone is not conclusive, since lava and glacial ice can carve channels too. So climb higher.

Rung two: minerals. Certain minerals only form in water. Orbital spectrometers mapped clays across ancient terrain and sulfates across younger terrain. In 2004 the Opportunity rover found small grey spheres nicknamed blueberries at Meridiani Planum, which proved to be hematite concretions, a form that on Earth grows in water-saturated rock, along with jarosite, a sulfate requiring acidic water. These are not shapes that might have another explanation; they are chemistry that requires liquid water.

Rung three: rocks in place. Curiosity landed in Gale Crater in 2012 and drove into an ancient lakebed at Yellowknife Bay. The mudstones there are fine-grained, laminated, and clay-bearing, exactly what settles out in still water, and their chemistry indicates neutral pH and low salinity, which is to say drinkable. Bed thickness and sedimentation rates suggest the lake system persisted, on and off, for thousands to millions of years. That is a habitable environment by any definition used on Earth.

Rung four: water today. Mars has water now, mostly frozen: a north polar cap of water ice up to 3 kilometres thick, a south cap with a carbon dioxide veneer over water ice, and vast ground ice at mid to high latitudes that Phoenix scraped down to in 2008. Two claims are contested. Recurring slope lineae, dark streaks lengthening down warm slopes in summer, were reported in 2015 as flowing brine but behave more like dry granular flows. And bright radar reflections beneath the south polar cap, reported in 2018 as a subglacial lake, may instead be clays or frozen brines.

Key idea: Landform shapes, water-requiring minerals mapped from orbit, and laminated lakebed mudstones sampled by a rover all converge on ancient standing water, and that convergence, not any single spectacular image, is what makes the conclusion robust.

Where the atmosphere went

A wet early Mars needs a thicker, warmer atmosphere than 6 millibars of carbon dioxide, and Mars receives only 586 watts per square metre against Earth's 1,361. So the climate problem has two halves: how was early Mars warm, and where did the air go?

The second half is better understood, and it is where the size argument from Module 2 pays off. Mars is small. It cooled fast, its dynamo shut down probably before 4.1 billion years ago, and volcanic resupply dwindled while escape processes kept working. Your calculation in Lesson 7 showed that ordinary thermal escape cannot remove oxygen from Mars, so the loss had to be non-thermal, and NASA's MAVEN orbiter, operating since September 2014, was built specifically to measure it.

MAVEN's findings are worth stating precisely because they are measurements rather than models. It sees ions escaping at a present-day rate on the order of a few kilograms per second, and has watched that rate jump by more than an order of magnitude during solar storms, which matters because the young Sun was far more active. The most elegant result, published by Bruce Jakosky and colleagues in 2017, uses argon. Argon is a noble gas: it cannot react or hide in rocks, so the only place it can go is space. Comparing light argon-36 with heavy argon-38, and knowing that sputtering preferentially removes the lighter isotope, they concluded that roughly two-thirds of Mars's original argon has been lost to space, implying Mars has lost the majority of the gas it once had.

The first half of the problem, how early Mars stayed warm enough for rivers under a Sun about 30 per cent fainter than today, remains genuinely unsolved. Carbon dioxide alone does not do it in most climate models, since the gas condenses out before delivering enough warming. Proposed supplements include hydrogen and methane, cirrus cloud warming, and episodic heating from impacts and volcanism, with some models producing not a persistently warm Mars but a cold one punctuated by warm intervals, which fits the geology reasonably well. File this faint young Sun problem as open.

One mechanism is uniquely Martian. With no large moon to stabilise it, Mars's axial tilt swings chaotically between roughly 0 and 60 degrees over millions of years, against Earth's tight 22 to 24.5 degree range, redistributing polar ice toward the equator at high obliquity and lurching the climate for purely orbital reasons.

The search for life, part one: Viking

The two Viking landers arrived in 1976 carrying three biology experiments and a gas chromatograph mass spectrometer, and the results were confusing in an instructive way.

The Labeled Release experiment added a radioactively tagged nutrient solution to Martian soil and watched for radioactive gas, which would indicate metabolism. It got a strong positive signal, and the signal vanished when the soil was heated first to sterilise it, which is also what you would expect from life. On its own it looked positive. But the gas chromatograph mass spectrometer found no organic molecules at all above its detection limits, and that was decisive for most scientists: you cannot have metabolism without organic chemistry. The consensus became that Martian soil contains strong chemical oxidants that decompose the nutrient abiotically, mimicking metabolism.

That explanation gained major support in 2008 when the Phoenix lander detected perchlorates in Martian soil at roughly half a per cent. Perchlorates are powerful oxidants, and when heated with organic compounds, as the Viking instrument did, they destroy those organics and produce chlorinated by-products. In hindsight Viking may have burned up any organics present before it could measure them. Gilbert Levin, the Labeled Release principal investigator, maintained until his death in 2021 that his experiment had detected life. He was a serious scientist making a serious argument, and the field did not agree with him; the dissent existed, and it did not persuade.

Case study: ALH84001

On 27 December 1984, a meteorite hunting team picked up a 1.93 kilogram greenish rock in the Allan Hills of Antarctica. Analysis later identified it as Martian, from trapped gases matching the Martian atmosphere. It is exceptionally old, with an igneous crystallisation age near 4.09 billion years, making it a sample of Noachian Mars, ejected roughly 16 million years ago and landing about 13,000 years ago.

On 16 August 1996, David McKay and colleagues published in Science a paper arguing that ALH84001 contained evidence of ancient Martian biological activity. The claim rested on four observations, all associated with carbonate globules in cracks in the rock.

1. Polycyclic aromatic hydrocarbons, or PAHs, organic molecules concentrated near the carbonates.
2. Magnetite and iron sulfide crystals whose size and shape resembled those produced by magnetotactic bacteria on Earth.
3. Carbonate globules with textures similar to some bacterially mediated carbonates.
4. Segmented ovoid structures, 20 to 100 nanometres long, described as possible nanofossils.

The announcement was enormous. President Clinton spoke about it from the White House lawn, and it contributed directly to the founding of the NASA Astrobiology Institute in 1998, creating the modern discipline.

Now the follow-up, which is the part worth studying.

On the PAHs: these are common in carbonaceous chondrites, form readily by abiotic chemistry, and occur in Antarctic meltwater, so contamination could not be excluded either. PAHs are not a biosignature.

On the magnetite: the argument was that certain crystal shapes are made only by bacteria. Others showed that heating iron-bearing carbonate produces similar magnetite, so impact heating works without biology, and the shape criteria proved less exclusive than claimed.

On the nanofossils: this line collapsed first. The structures are 20 to 100 nanometres across, while the smallest known free-living organisms are around 200 nanometres, which a 1999 US National Research Council panel identified as near the lower limit for a cell containing ribosomes and a genome. Objects hundreds of times smaller in volume cannot house known biochemistry, and some proved to be artefacts of the gold coating used for electron microscopy.

On the carbonates: a 2011 study led by Itay Halevy found the globules formed at about 18 degrees Celsius from evaporating brine. Low temperature is compatible with life but does not require it.

The most complete resolution came in 2022, when Andrew Steele and colleagues showed in Science that the organic material in ALH84001 was produced abiotically by water-rock reactions, specifically serpentinisation and carbonation. The organics are real, they are Martian, and they are not biological.

The current conclusion is clear: ALH84001 is not evidence of life on Mars. Every line of the original argument has a satisfactory abiotic explanation. But do not read this as a story about people being foolish. The original paper was carefully hedged and made testable claims that were then tested. The episode gave the field its most important methodological lesson, the requirement to exclude abiotic explanations before invoking biology, and it forced explicit standards of evidence for biosignature claims. It also confirmed something genuinely interesting: Mars has indigenous organic chemistry produced by water interacting with rock.

Key idea: ALH84001 failed as a biosignature because each of its four lines of evidence turned out to have an abiotic explanation, and the lasting result was a standard requiring abiotic routes be excluded before biology is invoked.

The search for life, part two: what is happening now

Curiosity has, since 2012, established that Gale Crater was habitable and detected indigenous organic molecules preserved in 3.5 billion year old mudstone, reported by Jennifer Eigenbrode and colleagues in 2018. Organics are not life; they are the raw material, and finding them preserved that long matters because the record was not entirely destroyed.

Curiosity has also measured methane: a background near 0.4 parts per billion that varies seasonally, with spikes an order of magnitude higher. Methane is destroyed by ultraviolet light within a few hundred years, so any present must be replenished, by geology or by biology. The complication is that ESA's far more sensitive ExoMars Trace Gas Orbiter has not detected it globally. Reconciling a rover that sees methane with an orbiter that does not has produced proposals including nighttime near-surface accumulation or localised seeps. As of 2026 it is unresolved.

Perseverance landed at the Jezero delta in February 2021, and its purpose is not to detect life directly but to collect and cache samples for return to Earth. In July 2024 the team reported a rock nicknamed Cheyava Falls with millimetre-scale leopard spot features associated with organic carbon and with iron phosphate and sulfide minerals. On Earth such redox patterns can be produced by microbes. They can also be produced without them, and the team said so plainly, stating that resolving it requires laboratory analysis on Earth.

The awkward part. Mars Sample Return has hit severe cost and schedule problems: the original architecture was judged unaffordable, alternatives were solicited, and as of early 2026 there is no confirmed architecture and no funded launch date. The samples sit in Jezero Crater, and this is a snapshot of a changing situation.

Common misconceptions

"Liquid water has been found flowing on Mars today." Recurring slope lineae were read that way in 2015, but the majority view now is dry granular flows, and the 2018 subglacial lake detection is contested. Mars has abundant frozen water and, at best, transient brines.

"Viking detected life and NASA covered it up." Viking produced one positive-looking result alongside a null organic detection, and the 2008 perchlorate discovery supplied a solid abiotic explanation. The dissent was public and published, not suppressed.

"Organic molecules on Mars mean life." Organic simply means carbon-based here. Carbonaceous meteorites are full of organics and abiotic water-rock reactions make them routinely, so organics are a prerequisite, not a result.

"Mars lost its atmosphere in one catastrophe." MAVEN measures ongoing loss today, and the argon evidence indicates a long, cumulative process.

Recap

Mars divides into the wet Noachian, the acidic Hesperian, and the cold dry Amazonian, recorded in clays, then sulfates, then anhydrous iron oxides. Ancient surface water is established by valley networks, outflow channels and deltas, by water-requiring minerals such as hematite concretions and jarosite, and by the laminated lakebed mudstones Curiosity examined at Gale. The atmosphere was lost gradually to non-thermal escape, measured by MAVEN, whose argon work indicates roughly two-thirds of that gas has gone to space, while how early Mars stayed warm under a fainter Sun remains open. The search for life runs from Viking's ambiguous results, explained by the 2008 perchlorate discovery, through ALH84001, whose four lines of evidence all acquired abiotic explanations, to Curiosity's preserved organics and unresolved methane and Perseverance's cached samples, whose return has no funded date as of early 2026.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Water and life on Mars. In Astronomy 2e (Section 10.5). OpenStax, Rice University. openstax.org
2. McKay, D. S., Gibson, E. K., Thomas-Keprta, K. L., et al. (1996). Search for past life on Mars: Possible relic biogenic activity in Martian meteorite ALH84001. Science, 273(5277), 924-930. doi.org
3. Steele, A., Benning, L. G., Wirth, R., et al. (2022). Organic synthesis associated with serpentinization and carbonation on early Mars. Science, 375(6577), 172-177. doi.org
4. Jakosky, B. M., Slipski, M., Benna, M., et al. (2017). Mars atmospheric escape and evolution from analysis of 36Ar/38Ar. Science, 355(6332), 1408-1410. doi.org
5. NASA Science. (n.d.). Mars exploration. National Aeronautics and Space Administration. science.nasa.gov
6. Encyclopaedia Britannica. (n.d.). Mars. britannica.com

The big picture

Add up the mass of every planet in the solar system and you get about 447 Earth masses. Jupiter alone is 318 of them, which is 71 per cent. Add Saturn at 95 and the two of them account for 92.5 per cent of all planetary material. Everything else, including Uranus, Neptune, and every rocky world, every moon, every asteroid and comet, shares the remaining 7.5 per cent. On a mass census, this is a solar system consisting of the Sun, Jupiter, Saturn, and some debris, and we live in the debris.

These planets are also fundamentally different objects from the ones in the last two modules. They have no surface. If you descended into Jupiter you would find the gas around you getting steadily denser and hotter until, without ever crossing a boundary you could stand on, you would be in a fluid, then in a strange metallic liquid, and then dead. The word planet is doing a lot of work covering both Mercury and Jupiter.

Two questions organise this lesson. What are these things made of and how do we know, given that we cannot see inside them or land on them? And why do two of them, Uranus and Neptune, remain so poorly understood in 2026 that a leading planetary scientist would call them the biggest gap in our knowledge of the solar system?

Planet	Mass (Earth = 1)	Density (g/cm^3)	Composition class	Heat emitted vs absorbed
Jupiter	318	1.33	Gas giant: mostly hydrogen and helium	About 1.7
Saturn	95.2	0.687	Gas giant: mostly hydrogen and helium	About 1.8
Uranus	14.5	1.27	Ice giant: mostly water, ammonia, methane	About 1.06
Neptune	17.1	1.64	Ice giant: mostly water, ammonia, methane	About 2.6

Two things in that table should stop you. Saturn's density of 0.687 g/cm^3 is less than water, so the old classroom line about Saturn floating in a big enough bathtub is arithmetically true, if physically absurd. And Uranus emits almost exactly as much heat as it absorbs while Neptune, farther out and colder, emits 2.6 times what it takes in. Uranus has an internal heat problem nobody has solved.

Key idea: Jupiter and Saturn hold 92.5 per cent of all planetary mass, they have no solid surface, and the gas giants and ice giants differ enough in composition that treating all four as one category hides more than it reveals.

Inside a gas giant

Descend into Jupiter and the pressure climbs relentlessly. At around one to two million times Earth's atmospheric pressure, roughly 80 per cent of the way out from the centre, hydrogen undergoes something remarkable: it stops behaving like a molecular gas and becomes metallic hydrogen, a liquid in which electrons are free to move as they do in a metal. This layer is enormous, it is electrically conducting, and Jupiter rotates once every 9 hours 56 minutes. Conducting fluid plus convection plus fast rotation is the dynamo recipe from Module 2, which is why Jupiter has the most powerful magnetic field of any planet.

At the centre there is a core, but the modern picture is stranger than the textbook one. The Juno orbiter, in Jovian orbit since July 2016, measured Jupiter's gravitational field with unprecedented precision, and the result does not fit a compact rocky core with a sharp boundary. It fits a dilute core, sometimes called fuzzy: heavy elements spread out and gradually mixed into the surrounding hydrogen over a substantial fraction of the planet's radius. The favoured explanation is a giant impact during formation that shattered and diluted an originally compact core, though the debate is open. Either way, this is a good example of a spacecraft measurement forcing revision of an idea that had been in textbooks for decades.

Saturn adds a twist. It emits about 1.8 times the energy it absorbs, more than simple slow contraction can supply. The leading explanation is helium rain: at Saturn's interior conditions helium becomes immiscible in metallic hydrogen, condenses into droplets, and falls toward the centre, releasing gravitational energy on the way. There is corroborating evidence, since Saturn's upper atmosphere is measurably depleted in helium relative to Jupiter's, which is what you would expect if helium has been draining downward.

The Galileo probe, dropped into Jupiter in December 1995, is worth knowing about as a lesson in sampling. It survived for 58 minutes and returned direct composition measurements, and it found far less water than models predicted. The reason turned out to be bad luck: it entered a so-called hot spot, a dry downwelling region roughly analogous to a desert, which is unrepresentative of the planet. Juno's microwave radiometer later measured water in Jupiter's equatorial region at roughly two to three times solar abundance, close to what had been expected. One probe at one point on a planet is one data point.

Weather on a world with no ground

Jupiter's banded appearance comes from alternating zones, bright bands of rising gas, and belts, darker bands of sinking gas, stretched into east-west stripes by rapid rotation. Between them run jet streams reaching hundreds of kilometres per hour. Juno discovered that these jets are not a shallow weather layer: they extend about 3,000 kilometres deep, involving a mass of atmosphere far greater than anyone had assumed.

The Great Red Spot is a high-pressure storm that has been observed for at least 190 years and possibly since the 1660s. It is shrinking. In the late 1800s it spanned roughly 40,000 kilometres; today it is around 14,000, or a bit more than one Earth diameter, and it has become rounder and taller. Nobody knows whether it is heading for dissolution or a new equilibrium.

Saturn's atmosphere is blander to the eye because a haze layer sits above its cloud decks, but it has its own oddity: a persistent, essentially perfect hexagon circling the north pole, about 30,000 kilometres across, which Voyager first saw and Cassini watched for years. Laboratory fluid experiments reproduce such polygons from a jet stream with the right shear, so it is unusual rather than mysterious.

Neptune, with the least sunlight of the four, has the fastest winds in the solar system, measured near 2,100 kilometres per hour. That is genuinely counterintuitive, since the energy driving weather has to come from somewhere, and it is part of why Neptune's large internal heat flux matters.

Magnetospheres

Jupiter's magnetic field is about 20,000 times stronger in total moment than Earth's, and the magnetosphere it inflates is the largest structure in the solar system apart from the Sun's own heliosphere. Its tail stretches beyond Saturn's orbit. If you could see it from Earth, it would appear several times larger than the full Moon.

Two consequences matter. First, Jupiter's volcanic moon Io injects about a tonne per second of sulfur and oxygen into a doughnut of plasma called the Io torus, which feeds the magnetosphere from the inside, an arrangement with no analogue at Earth. Second, the trapped particles form radiation belts so intense they are lethal to spacecraft electronics. Juno's most sensitive instruments sit inside a titanium vault, and Europa Clipper's design was driven substantially by radiation survival. Both Jupiter and Saturn also have permanent aurorae far brighter than Earth's, and Jupiter's are visible in ultraviolet from Earth orbit.

Key idea: Jupiter's dynamo runs on liquid metallic hydrogen rather than molten iron, and the resulting magnetosphere is both the largest structure in the solar system and a radiation hazard that shapes how missions to the Jovian system are designed.

The ice giants, and how little we know

Uranus and Neptune are not small Jupiters. Their bulk composition is dominated not by hydrogen and helium but by what planetary scientists confusingly call ices: water, ammonia and methane, which are not necessarily solid at the relevant conditions but which condensed as ices in the disk. A hydrogen-helium envelope wraps a hot, dense, electrically conducting fluid mantle. Laboratory work and simulations suggest water in there may exist as superionic ice, a phase with a rigid oxygen lattice through which hydrogen ions flow freely, behaving as both solid and liquid at once.

Their magnetic fields are bizarre. Instead of being roughly aligned with the rotation axis and centred on the planet, as at Earth, Jupiter and Saturn, they are tilted by about 59 degrees at Uranus and 47 at Neptune, and offset substantially from the centre. The likely explanation is that the dynamo runs in a relatively thin conducting shell rather than a deep core, but this is inference from two flybys.

And that is the problem. Every close-up measurement we have of Uranus and Neptune comes from a single spacecraft, Voyager 2, which flew past Uranus in January 1986 and Neptune in August 1989. Each encounter lasted days. We have no orbital data, no seasonal coverage, no probe measurements of composition, and no modern instruments applied to either world. Uranus is tipped 98 degrees onto its side, so its poles alternately point almost directly at the Sun through an 84 year orbit, producing seasons unlike anything else, and we have watched less than half of one.

This is why the 2022 US planetary science decadal survey ranked a Uranus Orbiter and Probe as its highest-priority large mission. Such a mission would take well over a decade to arrive. As of 2026 it has not been formally started, which is another snapshot to check rather than trust.

Worked example: the Roche limit and where rings live

All four giant planets have rings, and rings sit close in. The reason is the Roche limit: inside a certain distance, the tidal force stretching a large body apart exceeds its own self-gravity holding it together, so a moon cannot form or survive there and material stays as a swarm of particles.

For a fluid body held together only by gravity, the limit is

d = 2.44 x R_planet x (density of planet / density of moon)^(1/3)

Try it on Saturn. R = 60,268 km, Saturn's density is 0.687 g/cm^3, and an icy moon has a density of about 0.9 g/cm^3.

Density ratio = 0.687 / 0.9 = 0.7633. Its cube root is 0.914.

d = 2.44 x 60,268 x 0.914 = 147,054 x 0.914 = 134,400 km from Saturn's centre.

Saturn's bright main rings end at about 136,800 km. The prediction lands within two per cent of the observed outer edge of the main ring system, which for a one-line formula is a striking result. It tells you the rings are not an accident of history but a stable configuration: that region simply cannot hold a moon together.

Rings compared

Saturn's rings are unlike the others. They are 99.9 per cent water ice, bright and clean, spanning about 280,000 kilometres across but in most places only around ten metres thick. Scale that: if the main rings were the width of a football pitch, they would be far thinner than a sheet of paper.

Cassini's Grand Finale in 2017, when the spacecraft flew between the planet and the rings before being deliberately destroyed, allowed the ring mass to be measured directly from its gravitational pull. The answer, around 0.4 times the mass of the small moon Mimas, is low. Combined with how clean the ice is, since interplanetary dust should darken rings over time, this suggests the rings may be young, perhaps only 10 to 100 million years old, meaning they postdate the dinosaurs. That interpretation is contested, with some models allowing older rings that resist darkening, but it is a live and startling possibility.

The other systems are faint. Jupiter's rings are thin dust, replenished by micrometeorite impacts on the small moons Adrastea and Metis. Uranus has narrow, dark, tightly confined rings, discovered in 1977 not by a spacecraft but by watching Uranus pass in front of a star and seeing the starlight blink out and back several times before and after the planet itself. Neptune's rings include incomplete arcs, clumps held in place by gravitational shepherding. And in 2013 rings were found around Chariklo, an object only about 250 kilometres across, by the same occultation technique, which showed that rings are not the exclusive property of giant planets.

Common misconceptions

"Jupiter is a failed star." It would need roughly 80 times its present mass to fuse hydrogen. It is not a near miss; it is off by a factor approaching a hundred.

"The giant planets are all basically the same." Jupiter and Saturn are hydrogen and helium; Uranus and Neptune are dominated by water, ammonia and methane with only a thin hydrogen envelope. Their densities, heat balances and magnetic field geometries all differ.

"Saturn's rings are solid, or thick." They are countless independent ice particles from dust grains to boulders, and in most places the ring plane is only about ten metres thick.

"Uranus and Neptune are well studied." Every close measurement of either comes from one spacecraft passing by, once, in 1986 and 1989.

Recap

Jupiter and Saturn together hold 92.5 per cent of all planetary mass and are made mostly of hydrogen and helium, with a deep layer of liquid metallic hydrogen that powers Jupiter's dynamo; Juno's gravity data replaced the compact core with a dilute one. Saturn's excess heat is attributed to helium rain, supported by helium depletion in its upper atmosphere. Jupiter's jets run 3,000 kilometres deep, the Great Red Spot is shrinking, and Neptune has the fastest winds despite the least sunlight. Jupiter's magnetosphere is the largest structure in the solar system after the heliosphere, fed from within by the Io torus and dangerous enough to dictate spacecraft design. Uranus and Neptune are ice giants with wildly tilted, offset magnetic fields, and everything we know up close comes from two Voyager 2 flybys. Rings live inside the Roche limit, which for Saturn predicts an outer edge near 134,400 kilometres against an observed 136,800, and Cassini's mass measurement raises the real possibility that Saturn's rings are young.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). The giant planets. In Astronomy 2e (Section 11.2). OpenStax, Rice University. openstax.org
2. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Planetary rings. In Astronomy 2e (Section 12.5). OpenStax, Rice University. openstax.org
3. NASA Science. (n.d.). Jupiter. National Aeronautics and Space Administration. science.nasa.gov
4. NASA Science. (n.d.). Uranus. National Aeronautics and Space Administration. science.nasa.gov
5. Iess, L., Militzer, B., Kaspi, Y., et al. (2019). Measurement and implications of Saturn's gravity field and ring mass. Science, 364(6445), eaat2965. doi.org
6. National Academies of Sciences, Engineering, and Medicine. (2022). Origins, worlds, and life: A decadal strategy for planetary science and astrobiology 2023-2032. The National Academies Press. nap.nationalacademies.org

The big picture

Everything in this lesson is left over. Asteroids, comets, and Kuiper Belt objects are the material that never got incorporated into a planet, and that is exactly what makes them valuable. Planets melt, differentiate, and recycle themselves; they destroy their own histories. A small body that never got hot enough to melt has been sitting in cold storage for four and a half billion years holding the original recipe. If you want to know what the solar nebula was made of, you do not go to Earth. You go to a carbonaceous asteroid.

There is also a less academic reason to care. These objects cross our orbit, and occasionally one arrives. This is the only branch of astronomy with a civil defence application, and it is the only natural hazard that humanity could, in principle, prevent entirely. In 2022 we tested whether we can, and the test worked.

Along the way we will handle the Pluto question properly, because it is a good example of something students are often taught badly: a genuine scientific disagreement in which both sides have real arguments and the dispute is not about facts at all.

Key idea: Small bodies matter twice over, as the only unmelted record of what the solar nebula contained and as the one natural catastrophe that advance work could actually prevent.

Asteroids

Most asteroids orbit between Mars and Jupiter, roughly 2.1 to 3.3 AU. The belt is emptier than films suggest: the total mass of everything in it is about four per cent of the Moon, and roughly a third of that is Ceres alone. Typical separations between belt objects are millions of kilometres. Spacecraft cross it without any evasive manoeuvring at all.

Composition sorts by distance, preserving the frost line from Module 1. Inner-belt S-types are stony and dry. Outer-belt C-types are dark, carbon-rich, and often water-bearing, and they make up the majority. M-types appear metallic and are thought to be exposed cores of shattered differentiated bodies; NASA's Psyche mission, launched in October 2023 and due to arrive in 2029, is going to one to find out.

The Dawn mission visited the two largest. Vesta, 525 kilometres across, is differentiated, with a basaltic crust and an iron core, a genuine protoplanet that stopped growing. Ceres, 940 kilometres and round enough to be classified a dwarf planet, is quite different: roughly a quarter water ice by mass, with bright deposits of sodium carbonate in Occator Crater that were deposited by briny fluid reaching the surface geologically recently. Ceres has an interior that has been wet.

Close-up study of small asteroids produced a surprise. Itokawa, Bennu and Ryugu all turned out to be rubble piles: loose aggregates of fragments held together by weak gravity, with porosities of 30 to 50 per cent, rather than solid rocks. When OSIRIS-REx touched Bennu in 2020 to collect its sample, the surface behaved like a fluid and the spacecraft sank in much further than expected. This matters for planetary defence, because a rubble pile responds to being pushed very differently from a solid boulder.

Two sample returns have transformed asteroid science. Hayabusa2 brought back 5.4 grams of Ryugu in December 2020, and OSIRIS-REx delivered 121.6 grams of Bennu in September 2023. Analyses of both have found abundant organic material, hydrated minerals, and amino acids, including nucleobases, the components of RNA and DNA. This is important and easy to overstate: it confirms that the chemical building blocks of life were widely distributed in the early solar system and delivered to planetary surfaces by impacts. It says nothing whatsoever about whether life exists anywhere else. Building blocks are not buildings.

Comets and the Kuiper Belt

A comet nucleus is a few kilometres of ice and dust, and it does nothing at all until it comes close enough to the Sun for ices to sublimate. Then it grows a coma, a diffuse envelope that can exceed the diameter of the Sun, and two tails: a bluish ion tail pushed straight away from the Sun by the solar wind, and a curved yellowish dust tail that lags behind along the orbit. Both point generally away from the Sun, which means an outbound comet travels tail-first.

ESA's Rosetta mission orbited comet 67P/Churyumov-Gerasimenko from 2014 to 2016 and dropped the Philae lander onto it, the first landing on a comet. It found a startlingly dark, low-density, two-lobed object built from two pieces that merged gently, and it measured a deuterium-to-hydrogen ratio about three times Earth's, which as we saw in Lesson 7 argues against comets as the main source of Earth's water.

Comets come from two reservoirs. Short-period comets, returning in under 200 years and orbiting near the ecliptic plane, come from the Kuiper Belt beyond Neptune and the scattered disk. Long-period comets, arriving from any direction on enormous orbits, come from the Oort cloud, a spherical shell thought to extend from a few thousand to perhaps a hundred thousand astronomical units. It is worth being clear that the Oort cloud has never been observed. It is inferred entirely from the orbits of the comets that arrive from it, and that inference is strong, but you should file it as a well-supported model rather than a mapped structure.

The Kuiper Belt, by contrast, is directly observed, with thousands of catalogued objects between about 30 and 50 AU. New Horizons flew past Pluto in July 2015 and then past the small, cold, red, two-lobed object Arrokoth on 1 January 2019, giving us our only close look at a pristine, never-heated planetesimal. Its gentle two-lobe structure is direct physical evidence for slow accretion of the kind the streaming instability predicts.

A newer category is worth a mention: interstellar objects, passing through from other systems. Three have now been identified, beginning with 1I/Oumuamua in 2017 and 2I/Borisov in 2019, with 3I/ATLAS discovered in July 2025. They are rare, fast, and only observable briefly.

Key idea: Asteroid sample return has confirmed that amino acids and nucleobases were widespread in the early solar system and delivered to planets by impact, which establishes availability of ingredients and says nothing about whether life exists elsewhere.

The Pluto argument, presented fairly

Here is what actually happened. Pluto was discovered in 1930 and initially thought to be Earth-sized; its estimated mass fell repeatedly over the following decades, settling at about 0.2 per cent of Earth's. From 1992 onward, astronomers began finding many other objects in the same region. In 2005 Mike Brown's team found Eris, comparable in size to Pluto and about 27 per cent more massive.

That forced a decision, and it was a decision about words rather than about nature. Either Eris was a planet, and so were a growing list of similar objects, or Pluto was something else. In 2006 the International Astronomical Union adopted a definition requiring a planet to (a) orbit the Sun, (b) have enough gravity to pull itself into a round shape, and (c) have cleared the neighbourhood around its orbit. Pluto satisfies the first two and fails the third, since it shares its region with the Kuiper Belt population, so it was reclassified as a dwarf planet.

The case for the definition. Criterion (c) is not arbitrary. There is a measurable quantity capturing how thoroughly a body dominates its orbital zone, and when you compute it for solar system objects the results fall into two widely separated groups with a gap of several orders of magnitude between them. The eight planets are all in one group and every dwarf planet is in the other. That gap is a real physical feature of the solar system, reflecting whether a body won its accretion zone. Taxonomies that track real structure are good taxonomies, and without criterion (c) the planet count would keep growing indefinitely as surveys improve.

The case against. Alan Stern, who leads New Horizons, and others argue that the definition is bad on its own terms. It classifies an object by its surroundings rather than by what it is: move Earth to the Kuiper Belt and it would stop being a planet, which is a strange property for a category to have. Criterion (c) is also not crisply satisfied by the actual planets, since Earth shares its orbital space with thousands of near-Earth asteroids and Jupiter has its Trojans. Critics further note that the vote was taken on the final day of the assembly by a small fraction of the IAU's membership. They propose a geophysical definition instead: a planet is any body massive enough to be rounded by its own gravity and not undergoing fusion, regardless of where it orbits.

New Horizons then made the argument more pointed, because Pluto turned out to be geologically spectacular: water-ice mountains up to 3.5 kilometres tall, a vast convecting nitrogen-ice glacier in Sputnik Planitia with no impact craters on it at all, indicating it is being resurfaced today, layered atmospheric hazes, and evidence suggesting a subsurface liquid water ocean.

So who is right? Notice that nobody disagrees about a single fact. Both sides accept Pluto's size, mass, orbit, composition, and geology, and both accept that the eight major planets are dynamically dominant in a way Pluto is not. The disagreement is about which distinction a word should encode, which makes it a taxonomic dispute rather than an empirical one. That is a real kind of scientific argument, it is not resolvable by more data, and it is worth recognising when you meet one.

What hits us, and how often

Impactor size	Typical interval	Consequence	Example
A few metres	Several per year	Bright fireball, disintegrates high up	Routine
About 20 metres	Decades	Airburst, broken windows, injuries from glass	Chelyabinsk, 2013
About 50 to 60 metres	Centuries to a millennium	Airburst flattening a region	Tunguska, 1908
140 metres	Tens of thousands of years	Regional devastation, the survey threshold	None recorded
1 kilometre	Roughly every 500,000 years	Global climate effects	None in human history
10 kilometres	Roughly every 100 million years	Mass extinction	Chicxulub, 66 million years ago

The Chelyabinsk event on 15 February 2013 is the useful modern case. A roughly 20 metre object arrived undetected, exploded at about 30 kilometres altitude with an energy near 500 kilotonnes, and injured about 1,500 people, almost all from glass blown in by the shock wave. Nobody died. It arrived from the direction of the Sun, which is where our telescopes cannot look, and it is the reason space-based infrared survey telescopes matter.

Survey completeness as of the mid-2020s is roughly this: over 95 per cent of near-Earth asteroids larger than 1 kilometre have been found, and none of them is on a collision course this century. For the 140 metre class, the figure is closer to 40 to 45 per cent, which is the real gap. NASA's NEO Surveyor, an infrared space telescope designed to close it, is planned for launch around 2027. Treat both numbers as a snapshot.

Worked example: what DART actually did

On 26 September 2022 the DART spacecraft, about 570 kilograms, struck the small moon Dimorphos at 6.1 kilometres per second. Dimorphos, 160 metres across, orbits the larger asteroid Didymos, and that orbital period is easy to measure from Earth by timing eclipses, which is why the pair was chosen. Nothing in the system threatened Earth.

Start with the spacecraft's momentum:

p = m x v = 570 x 6,100 = 3.48 x 10^6 kg m/s.

Now the important physics. The impact did not merely stop the spacecraft; it blasted out a plume of ejecta, and that ejecta leaving the surface pushes back, like the exhaust from a rocket. The momentum enhancement factor, called beta, accounts for this, and DART's was measured at roughly 3.6, so the asteroid received about 3.6 times the momentum the spacecraft carried.

With Dimorphos's mass around 4.3 x 10^9 kg:

change in velocity = beta x p / m = 3.6 x (3.48 x 10^6) / (4.3 x 10^9) = 2.9 x 10^-3 m/s.

That is 2.9 millimetres per second, roughly walking pace divided by four hundred. And yet the orbital period of Dimorphos around Didymos fell from 11 hours 55 minutes to 11 hours 23 minutes, a change of about 32 minutes. The mission's success criterion had been 73 seconds. It beat that by a factor of twenty-six.

Two lessons. First, tiny velocity changes accumulate enormously over time, which is why warning time is the single most valuable resource in planetary defence: a nudge decades before a predicted impact does what no heroic effort can do at the last minute. Second, beta being well above 1 means ejecta did most of the work, and beta depends on the target's structure, which is why the rubble pile discovery matters and why ESA's Hera mission, launched in October 2024 and arriving at Didymos in late 2026, was sent to survey the crater and pin down Dimorphos's mass and internal structure.

Other techniques exist. A gravity tractor hovers nearby and tugs by mutual gravitation, which is slow but precise and works on any structure. Ion beam deflection pushes with a thruster plume. Nuclear options are a genuine last resort for short warning or very large objects, and they carry obvious risks including fragmenting the target. All of them, DART included, need years to decades of lead time.

One date to know: on 13 April 2029 the asteroid Apophis, about 340 metres across, will pass within roughly 31,600 kilometres of Earth's surface, closer than geostationary satellites and visible to the naked eye from parts of Europe and Africa. It will not hit us, then or for the rest of this century; that has been established by radar. NASA has redirected the OSIRIS-REx spacecraft, renamed OSIRIS-APEX, to study it during the encounter.

Common misconceptions

"The asteroid belt is a dangerous, crowded field." Its total mass is four per cent of the Moon and typical separations are millions of kilometres. Every spacecraft that has crossed it has done so without incident.

"Pluto was demoted because scientists made a mistake." Nothing about Pluto changed. Surveys found many similar objects, including the more massive Eris, forcing a choice about what the word planet should mean.

"A comet's tail streams behind it like a jet exhaust." Tails point away from the Sun regardless of the direction of travel, so an outbound comet leads with its nucleus and trails its tail ahead of it.

"Finding amino acids on an asteroid means life exists out there." It means the ingredients were common and got delivered to planets. Ingredients are necessary and nowhere near sufficient.

"We could deflect an asteroid discovered a few weeks out." Every viable technique, DART's kinetic impact included, produces a tiny velocity change that only becomes a large miss distance given years to decades of lead time.

Recap

Asteroids preserve unmelted nebular material, sorted by composition across the frost line, and returned samples from Ryugu and Bennu contain organics, hydrated minerals, amino acids and nucleobases, establishing that life's ingredients were widespread and delivered by impact. Small asteroids are rubble piles, which changes how they respond to being pushed. Comets come from the observed Kuiper Belt and the inferred Oort cloud, and Rosetta's deuterium measurement at 67P argues against comets as the main source of Earth's water. The Pluto reclassification followed the discovery of Eris and turns on whether a planet should be defined by dynamical dominance or by intrinsic roundness; both camps agree on every fact, making it a taxonomic rather than empirical dispute. Impact consequence scales steeply with size, over 95 per cent of kilometre-class near-Earth asteroids are catalogued against roughly 40 to 45 per cent of the 140 metre class, and DART changed Dimorphos's velocity by only 2.9 millimetres per second yet shortened its orbital period by 32 minutes, demonstrating deflection and showing that warning time is the decisive resource.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Asteroids and planetary defense. In Astronomy 2e (Section 13.2). OpenStax, Rice University. openstax.org
2. Thomas, C. A., Naidu, S. P., Scheirich, P., et al. (2023). Orbital period change of Dimorphos due to the DART kinetic impact. Nature, 616, 448-451. doi.org
3. Daly, R. T., Ernst, C. M., Barnouin, O. S., et al. (2023). Successful kinetic impact into an asteroid for planetary defence. Nature, 616, 443-447. doi.org
4. NASA Science. (n.d.). Planetary defense. National Aeronautics and Space Administration. science.nasa.gov
5. NASA Science. (n.d.). Pluto. National Aeronautics and Space Administration. science.nasa.gov
6. Encyclopaedia Britannica. (n.d.). Pluto. britannica.com

The big picture

For most of the twentieth century, the search for life beyond Earth meant the search for warm places, and warm meant close to the Sun. Mars was the candidate; everything past the asteroid belt was written off as frozen. That assumption was reasonable, and it was wrong.

We now think most of the liquid water in the solar system is not on Earth and not on Mars. It is hundreds of millions of kilometres from the Sun, sealed under tens of kilometres of ice, on moons that receive so little sunlight that surface temperatures sit below 130 kelvin. Europa alone probably holds twice as much liquid water as all of Earth's oceans combined. Add Enceladus, Ganymede, Callisto, Titan, and probably Triton and Pluto, and the outer solar system is wet in a way nobody expected before the 1980s.

That discovery reorganised astrobiology. If liquid water is the requirement, the habitable zone concept, which we will examine critically in the next lesson, points to a narrow band of orbits around a star. Subsurface oceans do not care about that band at all. They are heated from within, by tides, and they can exist essentially anywhere.

Key idea: Most of the solar system's liquid water is in subsurface oceans on outer moons heated by tides rather than sunlight, which decouples habitability from distance to the star.

Tidal heating: the engine

Gravity falls off with distance, so the near side of a moon is pulled harder toward its planet than the far side. That differential stretches the moon along the planet-moon line, raising a tidal bulge in the solid body itself, not just in any ocean.

If the orbit were a perfect circle, that bulge would be constant and nothing would happen. But if the orbit is eccentric, the moon's distance changes through each orbit, so the bulge grows and shrinks; and the moon's orbital speed varies while its rotation stays steady, so the bulge also rocks back and forth across the surface. The moon is being kneaded, and internal friction turns that flexing into heat.

Now the crucial subtlety. Tidal heating damps eccentricity. Left alone, a moon's orbit circularises and the heating shuts off within a geologically short time. So something must keep the orbit eccentric, and in the Jovian system that something is an orbital resonance: Io, Europa and Ganymede are locked in a 4:2:1 pattern, so that Io completes exactly four orbits for Europa's two and Ganymede's one. Their repeated aligned gravitational tugs pump each other's eccentricities and keep them from settling. Enceladus is similarly held in a 2:1 resonance with Dione.

The output is spectacular. Io, with no resonance partner beyond this arrangement and the strongest tides, is the most volcanically active body in the solar system, with hundreds of active volcanoes and plumes reaching 500 kilometres high, on a moon smaller than ours. Europa, farther out, gets less, which is enough to keep an ocean liquid without melting the surface.

Europa

Europa is slightly smaller than our Moon, with a surface of water ice so smooth that its total relief is measured in hundreds of metres, crossed by a bewildering network of dark ridges and bands. Crater counting gives a surface age of only about 40 to 90 million years, which means the surface is being renewed.

The strongest evidence for the ocean is magnetic, and it is worth understanding because it is a beautiful piece of indirect reasoning. Jupiter's magnetic field is tilted relative to its rotation axis, so as Jupiter spins, the field at Europa's location oscillates. A conducting layer inside Europa responds by generating its own induced magnetic field that opposes the change, and the Galileo spacecraft measured exactly such an induced field, with the right phase and amplitude. Ice is not a good conductor. Salty liquid water is. The measurement effectively requires a global layer of conducting fluid within about 100 kilometres of the surface.

Combine that with the geology, gravity data, and the fact that tidal heating supplies the necessary energy, and the standard picture is an ocean roughly 60 to 150 kilometres deep beneath an ice shell somewhere between about 15 and 25 kilometres thick. The ice thickness is genuinely uncertain and it matters enormously for any future mission that wants to get through it.

The chemistry is encouraging. Infrared spectra show salts on the surface, including sodium chloride identified in Hubble data in 2019, which suggests the ocean is in contact with a rocky sea floor rather than sealed between ice layers. Rock-water contact is what generates the chemical energy gradients life would need. Meanwhile Jupiter's radiation constantly processes the surface ice, producing oxidants such as hydrogen peroxide and molecular oxygen; if surface material cycles downward, that supplies the other half of a redox couple.

Plumes have been reported, from Hubble ultraviolet observations in 2012 and 2016 and from a 2018 reanalysis of old Galileo magnetometer and plasma data. These are suggestive and they are not confirmed. Unlike Enceladus, where the plumes are unmistakable, Europa's are at the edge of detectability, and you should treat them as an open question. Europa Clipper, launched on 14 October 2024 and arriving in April 2030, will make roughly fifty close flybys and should settle it.

Worked example: how much water is down there?

Europa's radius is 1,561 kilometres. Take an ice shell 20 kilometres thick and an ocean 100 kilometres deep beneath it. The ocean is then a spherical shell running from an inner radius of 1,441 km to an outer radius of 1,541 km.

Volume = (4/3) x pi x (R_outer^3 - R_inner^3)

1,541^3 = 3.659 x 10^9 and 1,441^3 = 2.992 x 10^9, so the difference is 6.67 x 10^8 km^3.

Volume = 4.189 x 6.67 x 10^8 = 2.79 x 10^9 km^3.

Earth's oceans hold about 1.335 x 10^9 km^3. So Europa's ocean is roughly 2.1 times the volume of every ocean on Earth, on a moon with about a quarter of Earth's diameter. The reason is geometric: Earth's water is a film averaging 3.7 kilometres deep on a large sphere, while Europa's is a shell a hundred kilometres deep on a small one. Redo the calculation with an ocean 60 kilometres deep and you get about 1.7 x 10^9 km^3, still more than Earth, which shows the conclusion is robust to the uncertainty.

Key idea: Europa's ocean is inferred principally from an induced magnetic field that requires a global conducting layer, and even conservative depth estimates make it larger than all of Earth's oceans combined.

Enceladus: the one that samples itself

Enceladus is tiny, only 504 kilometres across, and by every expectation of Module 2's size rule it should be a dead ball of ice. In 2005 Cassini found it venting.

From four warm fractures near the south pole, informally called the tiger stripes, roughly a hundred geysers jet water vapour and ice grains into space at hundreds of metres per second. Some of that material escapes and feeds Saturn's diffuse E ring. This is the single most convenient arrangement in astrobiology: a subsurface ocean that throws samples of itself into orbit, where a spacecraft can fly through and analyse them without landing or drilling. Cassini did exactly that, repeatedly.

What it found, over about a decade:

• Water, salty, with sodium and potassium salts at concentrations resembling Earth's oceans.
• Silica nanoparticles of a size and type that, on Earth, form only where water at above roughly 90 degrees Celsius interacts with rock. This is a strong argument for active hydrothermal systems on the ocean floor.
• Molecular hydrogen, at a few tenths of a per cent to over one per cent, reported by Hunter Waite and colleagues in 2017. Hydrogen is consumed quickly by chemistry, so a steady supply implies ongoing production, most plausibly by serpentinisation, water reacting with rock. On Earth, hydrogen plus carbon dioxide is a metabolism that whole communities of microbes run on.
• Organic molecules, including complex, large fragments.
• Phosphates, reported by Frank Postberg and colleagues in 2023, at concentrations far above Earth's seawater. This mattered because phosphorus was the last of the six major biogenic elements, carbon, hydrogen, nitrogen, oxygen, phosphorus and sulfur, not yet confirmed there.

Independently, Cassini measured Enceladus's physical libration, its slight rocking as it orbits, and found it too large for a rigid body, which requires the icy shell to be mechanically decoupled from the core by a global ocean, not just a regional south polar sea.

Sort this carefully. That Enceladus has a global salty ocean in contact with rock, containing organics, hydrogen and phosphates, with hydrothermal activity, is evidence and strong inference. That this constitutes a habitable environment by the standards we apply on Earth is a defensible inference. That anything lives there is speculation. One honest puzzle remains: Enceladus radiates more heat than steady-state tidal models comfortably produce, which may mean its activity is episodic, and an ocean that freezes and re-melts periodically is a harder place for anything to persist.

Titan: the other kind of ocean world

Titan is the only moon in the solar system with a substantial atmosphere: 1.5 bars at the surface, half again Earth's pressure, and 94 per cent nitrogen. That alone makes it unique. What makes it strange is what the remaining few per cent does.

At 94 kelvin, methane on Titan behaves the way water does on Earth. It evaporates, forms clouds, rains, carves river channels, and pools in lakes and seas, the largest of which, Kraken Mare, is bigger than the Caspian. These are liquid methane and ethane. Titan has a complete hydrological cycle running on hydrocarbons, and it is the only other place in the solar system with standing surface liquid.

Above all this, ultraviolet light and Saturn's magnetospheric particles break apart methane and nitrogen, and the fragments recombine into a haze of complex organic molecules called tholins, which rain down and pile into vast dune fields of organic sand around the equator. Titan is doing large-scale organic chemistry continuously.

ESA's Huygens probe descended through all of it on 14 January 2005 and landed on a damp plain of rounded, water-ice cobbles, the only landing ever made in the outer solar system. And beneath the surface, Cassini's gravity and rotation measurements indicate Titan also has a conventional subsurface water ocean, perhaps 50 to 80 kilometres down.

So Titan offers two possible habitats, and they deserve very different confidence levels. The deep water ocean is a fairly ordinary ocean-world target. The surface hydrocarbon lakes have been proposed as a home for an exotic biochemistry using methane as a solvent instead of water; this is a legitimate speculation that has generated real theoretical work, and it is speculation, with no supporting evidence. NASA's Dragonfly, a nuclear-powered rotorcraft that will fly between sites on Titan, is scheduled to launch in 2028 and arrive in the mid-2030s, a schedule that has already shifted more than once.

Triton, and the rest of the list

Neptune's largest moon Triton orbits backwards, opposite to Neptune's rotation. No moon that formed in place around its planet does that, so Triton is almost certainly a captured Kuiper Belt object, which makes it a sample of Pluto's population that we happened to get a close look at when Voyager 2 passed in 1989.

Its surface is 38 kelvin, the coldest directly measured anywhere in the solar system, and yet Voyager photographed dark plumes rising eight kilometres and drifting downwind, driven by nitrogen gas. Its surface has very few craters, so it is being renewed. Capture would have produced enormous tidal heating as the orbit circularised, and a subsurface ocean may survive.

The wider list of confirmed or probable ocean worlds now includes Ganymede, the largest moon in the solar system, which has both an induced field indicating an ocean and its own dynamo-generated field, the only moon that does; Callisto, with an induced field; Ceres, with brines reaching the surface; and possibly Pluto, where the nitrogen glacier of Sputnik Planitia and other features hint at liquid below. ESA's JUICE spacecraft, launched in April 2023, will enter orbit around Ganymede in the mid-2030s.

Ranked as astrobiological targets, the usual ordering is Enceladus first for accessibility, since it hands you samples, then Europa for the size and likely rock contact of its ocean, then Titan for its organic chemistry and its own ocean. The difficulty runs the other way: reaching any of these oceans directly means getting through kilometres of ice without contaminating what you find, which no one currently knows how to do.

Common misconceptions

"Ocean worlds are warm." Their surfaces are between about 38 and 130 kelvin. The oceans are liquid because tidal friction heats them from within and kilometres of ice insulate them, not because there is any warmth outside.

"Cassini found life on Enceladus." It found water, salts, silica indicating hydrothermal activity, hydrogen, organics, and phosphates. It found habitability indicators. It carried no life detection instrument.

"Titan's lakes are water." They are liquid methane and ethane at 94 kelvin. Titan's water is frozen as hard as rock and forms its bedrock, with a separate liquid water ocean far below.

"Europa's plumes are confirmed." They are reported from Hubble observations and a reanalysis of Galileo data, and they remain at the edge of detectability. Europa Clipper should resolve it after 2030.

Recap

Tidal flexing, sustained by orbital resonances that prevent eccentricity from damping away, heats the interiors of outer moons and keeps subsurface oceans liquid far outside any conventional habitable zone. Europa's ocean is established chiefly by the induced magnetic field Galileo measured, which requires a global conducting layer, and even conservative geometry gives it about twice the volume of Earth's oceans. Enceladus vents its ocean into space through south polar tiger stripes, and Cassini sampled water, salts, silica nanoparticles implying hydrothermal activity, molecular hydrogen, organics and phosphates, making it habitable by terrestrial standards without any evidence that it is inhabited. Titan runs a hydrological cycle on methane, builds tholins continuously, and hides a water ocean below. Triton is a captured Kuiper Belt object with nitrogen plumes at 38 kelvin. Ganymede, Callisto, Ceres and possibly Pluto extend the list, and every one of these oceans is behind kilometres of ice we do not yet know how to cross cleanly.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). The Galilean moons of Jupiter. In Astronomy 2e (Section 12.2). OpenStax, Rice University. openstax.org
2. Waite, J. H., Glein, C. R., Perryman, R. S., et al. (2017). Cassini finds molecular hydrogen in the Enceladus plume: Evidence for hydrothermal processes. Science, 356(6334), 155-159. doi.org
3. Postberg, F., Sekine, Y., Klenner, F., et al. (2023). Detection of phosphates originating from Enceladus's ocean. Nature, 618, 489-493. doi.org
4. NASA Science. (n.d.). Ocean worlds. National Aeronautics and Space Administration. science.nasa.gov
5. NASA Jet Propulsion Laboratory. (n.d.). Europa Clipper. California Institute of Technology. europa.nasa.gov
6. European Space Agency. (n.d.). Cassini-Huygens. esa.int

The big picture

Everything in this lesson rests on a single awkward fact, and it is worth stating as bluntly as possible before we start. We have exactly one example of life. Every organism ever studied, from the hyperthermophiles of hydrothermal vents to redwoods to you, belongs to one interrelated family using the same genetic code, the same twenty amino acids, the same handedness of sugars, and the same energy currency. Biology, as a science, is the study of a single case.

That means every statement in this lesson of the form life requires X is really a statement of the form the only life we know requires X. Sometimes that distinction does not matter much, because the physics or chemistry behind the requirement is general. Often it matters enormously. Keep the distinction visible as you read, because the whole discipline turns on it, and because the temptation to forget it is strongest exactly when a result is most exciting.

We will work through four requirements, then look at how far extremophiles stretch them, then face the origin of life honestly, and finally examine the habitable zone, the concept that organises most of the exoplanet search and that has more holes in it than its popularity suggests.

Key idea: Astrobiology reasons from a sample of one, so every stated requirement for life is a generalisation from terrestrial biochemistry rather than a law, and the strength of each requirement depends on how general the underlying physics is.

What even counts as life?

There is no agreed definition. The most widely used working formulation, adopted by NASA for practical purposes, is that life is a self-sustaining chemical system capable of Darwinian evolution. It is serviceable and it leaks at the edges: viruses do not self-sustain but do evolve, a fire sustains itself but does not evolve, and a mule is unambiguously alive and cannot reproduce at all.

This is not pedantry, because a definition determines what an instrument looks for. Design a detector around metabolism and you will miss dormant spores; design it around replication and you must watch for generations. Viking's difficulties were partly a definitional problem wearing a hardware costume.

Requirement one: energy

Life is thermodynamically expensive. It builds and maintains order, which requires a continuous flow of usable energy, and what matters is not heat but a gradient, a disequilibrium that can be tapped. A uniformly warm environment at equilibrium is useless however warm it is.

On Earth, sunlight powers most of the biosphere. But entire ecosystems run without it. At deep-sea hydrothermal vents, chemosynthetic bacteria harvest chemical energy from reactions between vent fluids and seawater, and whole communities of tube worms and crabs depend on them. That 1977 discovery at the Galapagos Rift removed sunlight from the list of necessities. It is why Enceladus's molecular hydrogen matters: hydrogen plus carbon dioxide is a real metabolism, and Enceladus supplies both.

Requirement two: a solvent

Chemistry needs a medium in which molecules can move and meet. Water is extraordinarily good at the job, for reasons worth listing because they are chemistry, not chauvinism.

• It is polar, so it dissolves an enormous range of compounds, which is why it is called the universal solvent.
• It has a wide liquid range, 100 kelvin at Earth's surface pressure.
• It has an unusually high heat capacity, so watery environments buffer temperature swings.
• Solid water floats, an anomaly caused by hydrogen bonding, so bodies of water freeze from the top down and preserve liquid beneath. Almost every other substance would freeze solid from the bottom up.
• It is abundant, made from the first and third most common chemically active elements in the universe.

Alternatives deserve fair treatment. Ammonia is polar and liquid from 195 to 240 K at one bar, suiting colder worlds, but its liquid range is narrower and it dissolves the molecules we know about less well. Liquid methane, as on Titan, is non-polar, so it dissolves entirely different compounds, and it is cold enough that reactions run extremely slowly. Formamide and supercritical carbon dioxide have advocates. The honest summary is that water is demonstrably sufficient, the alternatives are speculative, and astrobiology follows the water not because water is proven necessary but because it is the only case we can reason about with evidence.

Requirement three: chemistry

Terrestrial life is built from six elements, CHNOPS: carbon, hydrogen, nitrogen, oxygen, phosphorus and sulfur. Carbon does the structural work, and it is worth understanding why it is favoured on general grounds rather than parochial ones. Carbon forms four covalent bonds, so it can build chains, rings and branched networks; those bonds are strong enough to be stable at biological temperatures but weak enough to be broken and remade by enzymes; and it bonds happily with hydrogen, oxygen, nitrogen and sulfur, giving an enormous space of possible molecules.

Silicon sits directly below carbon and also forms four bonds, which is why it appears in science fiction. In practice it is a poor substitute. Silicon-silicon bonds are much weaker than carbon-carbon bonds, so long silicon backbones are unstable. Silicon does not readily form double or triple bonds. And the decisive practical problem: carbon's fully oxidised form, carbon dioxide, is a gas that dissolves and circulates freely, whereas silicon's, silicon dioxide, is quartz, a rock. A silicon-based metabolism would clog itself with sand. The universe seems to agree, since interstellar clouds and meteorites are full of complex carbon chemistry and contain nothing comparable in silicon.

Key idea: Carbon and water are favoured on chemical grounds that hold anywhere in the universe, not merely because they are what we happen to be made of, which is why following the water is a defensible strategy and not just a failure of imagination.

Requirement four: time and stability

On Earth, life appeared early. The oldest widely accepted evidence is about 3.5 billion years old, stromatolites in the Dresser Formation of Western Australia, with isotopically light carbon in Greenland's Isua rocks pushing a contested case to roughly 3.7 billion years. Earth itself became habitable perhaps 4.3 billion years ago, so life may have taken only a few hundred million years to appear, which is fast.

People draw a conclusion from that speed: if it happened quickly here, it must be easy, so it should be common. Be careful. We could only be having this conversation on a planet where life did arise, so the sample is selected by its own outcome. This observation selection effect means a rapid origin on Earth is compatible both with life being easy and with it being fantastically rare, and one data point cannot separate those.

What the record does show is that complexity was slow. Simple cells persisted for roughly two billion years before the first eukaryotic cells, and animals appear only in the last 600 million years or so. If that ordering is typical, microbial life could be widespread while complex life is scarce.

Extremophiles: what they actually widen

Since the 1960s, organisms have been found thriving in conditions once assumed sterilising.

Type	Condition tolerated	Example
Hyperthermophile	Up to about 122 degrees Celsius under pressure	Methanopyrus kandleri
Psychrophile	Active down to about -20 degrees Celsius in brine veins	Antarctic sea ice communities
Halophile	Saturated salt solutions	Halobacterium in the Dead Sea
Acidophile	pH near 0	Picrophilus
Piezophile	Over 1,000 bar at full ocean depth	Mariana Trench microbes
Radioresistant	Thousands of grays, a dose that kills a human at about 5	Deinococcus radiodurans
Endolith	Inside rock kilometres below the surface	Deep continental subsurface communities

The deep subsurface case is the most astrobiologically important. Microbes live in fractured rock several kilometres down, metabolising so slowly that individual cells may divide once in centuries. This deep biosphere holds a substantial share of all life on Earth, and if Mars retains life, this is almost certainly where.

Now the part that gets left out. Every extremophile ever found still uses liquid water, DNA and RNA and proteins, the same twenty amino acids, and ATP. They have evolved remarkable protective machinery: heat-stable proteins, specialised membranes, extraordinary DNA repair. They have not evolved different chemistry. Extremophiles widen the range of conditions under which the one known biochemistry operates, and provide no evidence for alternative biochemistries.

A cautionary episode makes the point. In 2010 a NASA-funded team reported a Mono Lake bacterium, GFAJ-1, said to substitute arsenic for phosphorus in its DNA, which would have been genuinely different biochemistry. It was announced at a press conference and immediately criticised on methodological grounds. By 2012 two independent studies showed the organism is an unusually arsenic-tolerant extremophile that still requires phosphorus. That is a second case, alongside ALH84001 and phosphine, of the same lesson.

The origin of life: unsolved, and honestly so

This is the biggest gap in the field, and there is no point dressing it up. Nobody has made life from non-life, and nobody knows how it happened.

Progress has been real. Stanley Miller's 1953 experiment, running electrical discharges through methane, ammonia, hydrogen and water, produced amino acids in days, establishing that biological building blocks form readily from simple chemistry. Its weakness is that the early Earth's atmosphere was probably not that strongly reducing. That weakness has been patched from elsewhere: as Module 4 showed, amino acids and nucleobases arrive ready-made in carbonaceous meteorites, so availability of building blocks is not the problem.

The problem is the next step. The leading framework is the RNA world, which solves a chicken-and-egg puzzle: DNA stores information but cannot copy itself without proteins, and proteins are built from DNA instructions. RNA can do both, storing information and catalysing reactions as a ribozyme. So perhaps RNA came first. The difficulties are severe: ribose is hard to make and unstable, assembling RNA chains without enzymes is hard, and no one has demonstrated an RNA molecule that copies itself indefinitely without help.

A rival family puts metabolism first, with self-sustaining reaction networks on mineral surfaces preceding any genetic molecule. Alkaline hydrothermal vents such as the Lost City field are a favoured setting, providing natural pH and redox gradients across thin mineral membranes that resemble the chemiosmotic gradients all cells use. Others favour Darwin's warm little pond, with wet-dry cycling concentrating and polymerising molecules. Jack Szostak's group has built protocells, fatty acid vesicles that grow, divide, and admit nucleotides.

Every one of these is a piece. None of them, nor all together, gets from chemistry to a self-replicating evolving system. Panspermia relocates the origin without explaining it.

Here is why this matters more than any other gap in the course. Because we do not know the mechanism, we cannot estimate the probability. Not roughly, not to within orders of magnitude. Anyone who tells you life is probably common, or probably vanishingly rare, is expressing a preference. Hold that thought for the Drake equation in Module 6.

Key idea: The transition from chemistry to biology has never been demonstrated, so the probability of life arising is unknown by orders of magnitude, and every confident claim about how common life is rests on intuition rather than evidence.

The habitable zone, and its limits

The habitable zone is the range of distances from a star where a rocky planet with an Earth-like atmosphere could sustain liquid water on its surface. The inner edge is set by a runaway or moist greenhouse, the outer by carbon dioxide condensing out of the atmosphere faster than the greenhouse can compensate. For the Sun, the conservative estimates from James Kasting's 1993 work and Ravi Kopparapu's 2013 refinement put it roughly between 0.95 and 1.37 astronomical units.

Scaling to another star is simple, because what matters is the flux received, which depends on the star's luminosity:

distance = solar-system boundary x sqrt(L / L_sun)

Worked example. Proxima Centauri has a luminosity of about 0.0017 times the Sun's. Then sqrt(0.0017) = 0.0412, so

inner edge = 0.95 x 0.0412 = 0.039 AU, outer edge = 1.37 x 0.0412 = 0.056 AU.

Proxima b orbits at 0.0485 AU, comfortably inside that band, which is why it gets called potentially habitable. But look at what those numbers mean. The whole habitable zone is 0.017 AU wide, about 2.5 million kilometres, and it sits closer to the star than Mercury is to the Sun. A planet there completes an orbit in 11 days and is almost certainly tidally locked, one face permanently lit. Proxima is also a flare star that emits enormous ultraviolet and X-ray outbursts, and it spent its first few hundred million years far more luminous than it is now, which may have stripped the planet's atmosphere before things settled down.

So the concept has serious limitations you should be able to list.

1. It considers only surface liquid water. Every ocean world in the previous lesson lies far outside any habitable zone, and between them they hold most of the solar system's liquid water.
2. It assumes an Earth-like carbon dioxide and water atmosphere with a working carbonate-silicate cycle. Change the atmosphere and the boundaries move.
3. It ignores planet mass, magnetic field, composition, obliquity, and history.
4. Around small stars it delivers tidal locking, flares, and an early luminous phase, none of which the calculation sees.
5. Most simply: Venus is at the edge of the Sun's habitable zone. A planet being in the zone tells you it is worth looking at, not that it is habitable.

Hold it as a target-selection heuristic. It answers where to point the telescope first, not which planets have life.

Common misconceptions

"Extremophiles show life can use different chemistry." They show the opposite: every one uses water, nucleic acids, proteins and ATP, widening conditions not biochemistry.

"Miller and Urey showed how life began." They showed that amino acids form easily from simple molecules. The gap between amino acids and a self-replicating evolving system is where the entire problem lives.

"Life arose quickly on Earth, so it must be common." We could only observe this from a planet where it did arise, so the observation is selected by its own outcome and cannot distinguish easy from vanishingly rare.

"A planet in the habitable zone is habitable." Venus sits near the inner edge of the Sun's. The zone is a rough flux criterion, not a verdict.

"Silicon-based life is a serious alternative." Silicon backbones are unstable and silicon dioxide is a rock, so such a metabolism could not move its waste.

Recap

Life as we know it needs an energy gradient, a solvent, a versatile chemistry, and time, each generalised from a single example. Water and carbon are favoured for reasons of physics and chemistry rather than familiarity, which makes following the water defensible; ammonia and liquid methane remain speculative alternatives. Extremophiles have widened the range of temperature, pressure, salinity, acidity and radiation under which life persists, including a deep subsurface biosphere that is the most likely refuge on Mars, but every one uses the same biochemistry, and the GFAJ-1 arsenic episode shows what happens to claims otherwise. The origin of life remains unsolved: building blocks are abundant, the RNA world and metabolism-first frameworks each explain part, and no one has crossed from chemistry to biology, which is why the probability of life arising cannot be estimated. The habitable zone, scaled by the square root of luminosity, is a target-selection tool that ignores subsurface oceans, atmospheric variety, stellar activity, and the fact that Venus lies within the Sun's.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Astrobiology. In Astronomy 2e (Section 30.2). OpenStax, Rice University. openstax.org
2. NASA Astrobiology. (n.d.). About astrobiology. National Aeronautics and Space Administration. astrobiology.nasa.gov
3. Kopparapu, R. K., Ramirez, R., Kasting, J. F., et al. (2013). Habitable zones around main-sequence stars: New estimates. The Astrophysical Journal, 765(2), 131. doi.org
4. Reaves, M. L., Sinha, S., Rabinowitz, J. D., Kruglyak, L., & Redfield, R. J. (2012). Absence of detectable arsenate in DNA from arsenate-grown GFAJ-1 cells. Science, 337(6093), 470-473. doi.org
5. National Oceanic and Atmospheric Administration. (n.d.). Hydrothermal vents. Ocean Exploration. oceanexplorer.noaa.gov
6. Encyclopaedia Britannica. (n.d.). Extraterrestrial life. britannica.com

The big picture

In 1994 the number of known planets orbiting normal stars other than the Sun was zero. By early 2026 the confirmed count is roughly 6,000, and it climbs every few weeks as new candidates are validated. Treat that number as a snapshot; the NASA Exoplanet Archive keeps the running total, and any figure printed in a course is out of date by the time you read it.

The interesting question is not how many but how. Planets do not shine. They are small, they sit next to something overwhelmingly brighter, and they are far away. Almost the entire field is built on a handful of clever indirect techniques that detect a planet by its effect on the star rather than by seeing it at all, and every one of those techniques finds a particular kind of planet and misses others. Understanding the biases is not a footnote; it is the only way to read the population statistics in the next lesson without being misled.

This lesson is the quantitative core of the module. You will work the two dominant methods numerically, with real systems, so that when you read that a planet has a radius of 1.3 Earths and a mass of 4 Earths, you know exactly which measurement produced which number.

Key idea: Nearly all exoplanets are detected indirectly through their effect on their star, and each method has a distinct selection bias that shapes what the catalogue contains.

Why not just take a picture?

Two numbers explain the difficulty. First, contrast. In visible light the Sun outshines Earth by a factor of about 10^10, ten billion. Detecting Earth beside the Sun is like spotting a firefly next to a searchlight. Second, angular separation. Seen from 10 parsecs, about 33 light years, the Earth-Sun separation of one astronomical unit subtends 0.1 arcseconds, which is a tenth of the width of the smallest detail typical ground-based telescopes resolve through the atmosphere.

Direct imaging does work, but only in a specific corner of parameter space: young, massive, self-luminous planets on wide orbits, observed in the infrared where a hot young giant is far brighter relative to its star. The HR 8799 system, with four imaged giant planets, is the classic case. Those planets are millions rather than billions of years old and still glowing from formation heat. Nothing resembling Earth has ever been directly imaged.

The radial velocity method

A star and planet both orbit their common centre of mass, so the star traces a small circle and its velocity toward and away from us varies periodically. That shifts its spectral lines by the Doppler effect, and modern spectrographs can measure such shifts to better than a metre per second.

For a circular orbit, the star's velocity amplitude is well approximated by

K = 28.4 m/s x (M_p / M_Jupiter) x (M_star / M_Sun)^(-2/3) x (P / 1 year)^(-1/3)

Jupiter, seen from outside. M_p = 1 Jupiter mass, M_star = 1 solar mass, P = 11.86 years.

(11.86)^(-1/3) = 1 / 2.281 = 0.4384, so K = 28.4 x 1 x 1 x 0.4384 = 12.4 m/s.

Jupiter makes the Sun move at about 12 metres per second, roughly the speed of a fast cyclist. That is detectable.

Earth, seen from outside. Earth is 1/317.8 of a Jupiter mass, which is 0.003147, and P = 1 year.

K = 28.4 x 0.003147 x 1 x 1 = 0.089 m/s, which is 8.9 centimetres per second.

That is walking pace divided by fifteen, measured across tens of light years, on a star whose surface is churning with convection that produces its own velocity noise at the metre-per-second level. Detecting an Earth analogue by radial velocity is at or beyond the current state of the art, and stellar activity, not instrument precision, is the limiting problem.

51 Pegasi b, the first one. M_p sin i = 0.46 Jupiter masses, M_star = 1.11 solar masses, P = 4.23 days = 0.0116 years.

(1.11)^(-2/3) = 0.933 and (0.0116)^(-1/3) = 1/0.2262 = 4.42.

K = 28.4 x 0.46 x 0.933 x 4.42 = 54 m/s.

The measured value is about 56 metres per second. A Jupiter-mass planet whipping around its star every four days produces a signal four times larger than our Jupiter's, which is exactly why the first discovery was a hot Jupiter and not something familiar. Note also the sin i in that mass: radial velocity measures only the component of motion along our line of sight, so on its own it yields a minimum mass, not a true mass.

Key idea: The radial-velocity amplitude scales with planet mass and inversely with the cube root of the period, so short-period massive planets produce the largest signals, which is why they were found first.

The transit method

If a planet's orbit happens to be edge-on as seen from Earth, the planet passes in front of its star once per orbit and blocks a little light. The fractional dip is simply the ratio of the areas:

depth = (R_planet / R_star)^2

Jupiter across the Sun. R_Jupiter = 69,911 km, R_Sun = 696,000 km.

ratio = 0.1004, so depth = 0.01008, about 1.0 per cent, or 10,000 parts per million. Easily measured, even from the ground.

Earth across the Sun. R_Earth = 6,371 km.

ratio = 0.009154, so depth = 8.38 x 10^-5, which is 0.0084 per cent, or 84 parts per million.

Eighty-four parts per million is why Kepler had to go to space. Earth's atmosphere alone introduces variability far larger than that.

Earth across a small red dwarf. Now take a star of 0.2 solar radii, so 139,200 km.

ratio = 6,371 / 139,200 = 0.04577, so depth = 2.10 x 10^-3, which is 0.21 per cent, or 2,095 parts per million.

The same planet produces a signal 25 times deeper around the small star. This single fact explains a great deal about modern exoplanet science: M dwarfs are targeted heavily not because anyone believes they are the best homes for life, but because they are the only stars around which small planets are currently detectable and characterisable.

Transits have a second, harsher requirement: the geometry has to cooperate. The probability that a randomly oriented orbit transits is approximately

probability = R_star / a

For Earth: 696,000 / 1.496 x 10^8 = 0.0047, so 0.47 per cent. Fewer than one in two hundred randomly oriented Earth-Sun systems would show a transit at all. For a hot Jupiter at 0.05 AU: 696,000 / 7.48 x 10^6 = 0.093, about 9.3 per cent, twenty times better. Transit surveys therefore have to watch enormous numbers of stars, which is precisely what Kepler did: it stared at about 150,000 stars continuously for four years.

Putting the two together

The methods are complementary in a way that matters enormously. A transit gives you the planet's radius. A radial velocity measurement gives you its mass, and because a transit tells you the orbit is nearly edge-on, sin i is close to 1 and the minimum mass becomes a true mass. With both you get

density = mass / volume

and density is what tells you whether an object is rock, ice, or gas. A planet of 1.5 Earth radii and 5 Earth masses has a density around 8 g/cm^3 and is a dense rock. A planet of the same radius and 2 Earth masses has a density near 3.3 and probably carries a thick atmosphere or a substantial water layer. Without both measurements, you have half a planet.

The other methods, and what each one misses

Microlensing uses general relativity: when one star passes precisely in front of a more distant one, its gravity focuses the background starlight, and a planet orbiting the foreground star adds a brief extra spike. It is uniquely sensitive to planets at intermediate to wide separations, and to free-floating planets bound to no star at all. Its drawback is severe: the alignment never repeats, so a microlensing planet can never be re-observed.

Astrometry measures the star's tiny side-to-side wobble on the sky rather than along the line of sight. It favours massive planets on wide orbits around nearby stars, and ESA's Gaia mission is expected to deliver a large astrometric planet catalogue.

Transit timing variations detect additional planets by the gravitational tugs they exert on a known transiting planet, shifting its transits earlier or later. This is how much of the TRAPPIST-1 system was weighed.

Pulsar timing deserves a historical note: the very first confirmed exoplanets, found by Aleksander Wolszczan and Dale Frail in 1992, orbit a pulsar. Pulsar timing is extraordinarily precise but applies to a rare and hostile class of star.

Now the biases, stated plainly, because the next lesson depends on them.

Method	Measures	Favours	Systematically misses
Radial velocity	Minimum mass, period, eccentricity	Massive planets on short periods	Small planets, long periods, active stars
Transit	Radius, period, atmosphere access	Large planets, short periods, small stars	Non-aligned orbits, which is over 99 per cent of Earth analogues
Direct imaging	Brightness, spectrum, orbit	Young, hot, massive, wide-separation planets	Everything old, small, or close in
Microlensing	Mass ratio, separation	Wide separations, free-floating planets	Repeat observation of anything
Astrometry	True mass, orbit	Massive, wide, nearby	Close-in and low-mass planets

Key idea: A transit gives radius and a radial velocity gives mass, and only together do they give density, which is the number that says whether a planet is rock, water, or gas.

The survey machines

Kepler operated from 2009 to 2018, staring at a single patch of sky and then, after two of its reaction wheels failed, at a series of fields along the ecliptic in its extended K2 mission. It confirmed over 2,700 planets and, more importantly, delivered the statistically well-characterised sample that made occurrence rates possible. TESS, launched in 2018, surveys almost the whole sky in shorter stares, finding planets around bright nearby stars that are good follow-up targets. CHEOPS refines radii of known planets, Gaia supplies astrometry and precise stellar parameters, and ESA's PLATO, designed to find small planets on longer periods around bright Sun-like stars, is the next major step.

Common misconceptions

"We photograph exoplanets routinely." Only a few dozen have been directly imaged, all young, massive, and far from their stars. The rest are inferred from starlight.

"Transit surveys see most planets." They see only the small fraction whose orbits are aligned with our line of sight, which for an Earth-Sun geometry is about one in two hundred.

"Radial velocity gives a planet's mass." It gives a minimum mass, mass times sin i, unless the inclination is known independently, which for transiting planets it is.

"Red dwarfs are targeted because they are best for life." They are targeted because a given planet produces a much deeper transit and a larger stellar wobble around a small star, making detection and follow-up feasible.

Recap

Direct imaging is defeated by a contrast of about 10^10 and a separation of 0.1 arcseconds for an Earth analogue at 10 parsecs, so it works only for young, hot, wide giants. Radial velocity measures the star's Doppler wobble, with amplitude 28.4 m/s scaled by planet mass and the inverse cube root of the period, giving 12.4 m/s for Jupiter, 8.9 cm/s for Earth, and 54 m/s for 51 Pegasi b, and it yields only a minimum mass. Transits give a depth equal to the square of the radius ratio, so Jupiter across the Sun is 1 per cent, Earth across the Sun is 84 parts per million, and Earth across a 0.2 solar radius star is 2,095 parts per million, twenty-five times deeper, which is why M dwarfs dominate the target lists. Transit probability is roughly the stellar radius over the orbital distance, only 0.47 per cent for an Earth analogue, which is why surveys watch hundreds of thousands of stars. Radius from transit plus mass from radial velocity gives density, and density is what distinguishes rock from gas.

Sources

1. NASA Exoplanet Archive. (n.d.). Confirmed planets counts and data. NASA Exoplanet Science Institute, Caltech. exoplanetarchive.ipac.caltech.edu
2. NASA Science. (n.d.). Exoplanets: Ways to find a planet. National Aeronautics and Space Administration. science.nasa.gov
3. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Planets beyond the solar system: Search and discovery. In Astronomy 2e (Section 21.5). OpenStax, Rice University. openstax.org
4. Mayor, M., & Queloz, D. (1995). A Jupiter-mass companion to a solar-type star. Nature, 378, 355-359. doi.org
5. NASA Ames Research Center. (n.d.). Kepler and K2 mission. National Aeronautics and Space Administration. nasa.gov

The big picture

Before 1995 there was one planetary system to reason from, and every theory of planet formation was built to reproduce it. Small rocky worlds close in, gas giants beyond the frost line, everything on near-circular orbits in the same plane. It was assumed to be typical because there was nothing to compare it with.

Roughly 6,000 planets later, that assumption has not survived. The commonest kind of planet in the galaxy is a size that does not exist in our solar system at all. Giant planets turn up on four-day orbits. Systems exist with seven planets packed inside the orbit of Mercury. And the configuration we live in, small rocks inside, giants far outside, on a long-period, well-spaced arrangement, has not obviously been found anywhere else.

The honest question is whether that last fact means anything. It might mean the solar system is unusual. It might mean our methods are worst at finding systems like ours. Distinguishing those two is the central interpretive problem of this lesson, and it is a case study in reading data through the biases you learned to compute in the last one.

Key idea: The exoplanet catalogue is not a random sample of planetary systems, so every statement about what is typical must be corrected for detection bias before it means anything.

What the catalogue contains

The population sorts roughly into these groups. Bear in mind that names describe size and location, not necessarily composition.

Class	Rough size	Notes
Hot Jupiter	Giant, orbiting in under about 10 days	Rare, around 0.5 to 1 per cent of Sun-like stars, but heavily overrepresented in early catalogues
Warm and cold giants	Giant, wider orbits	Harder to find, need long baselines
Mini-Neptune	About 2 to 4 Earth radii	Rocky or icy core with a thick hydrogen and helium envelope
Super-Earth	About 1 to 1.7 Earth radii	Probably rocky, larger than Earth, no solar system analogue
Terrestrial	Under about 1 Earth radius	Detectable mainly around small stars

Two facts stand out. First, super-Earths and mini-Neptunes together are the most common type of planet we find, and our solar system contains none. There is a gap between Earth at 1.0 Earth radii and Neptune at 3.9, and most of the galaxy's planets appear to live in it. Any formation theory has to explain both why they are so common elsewhere and why we have none.

Second, there is structure inside that range. In 2017 Benjamin Fulton and colleagues, working with precisely measured Kepler planets, found a statistically significant deficit of planets between about 1.5 and 2.0 Earth radii. This is the radius valley, or Fulton gap, and it splits the small-planet population into two peaks.

The explanation is atmospheric loss, and it connects directly to Module 3. A small planet close to its star either keeps a hydrogen and helium envelope or loses it. Keep it and even a thin envelope inflates the measured radius above 2 Earth radii. Lose it and you are left with the bare rocky core below 1.5. The intermediate sizes are unstable, so few planets are found there. Two mechanisms compete for the credit: photoevaporation, in which the young star's extreme-ultraviolet output drives the envelope off, and core-powered mass loss, in which the planet's own residual formation heat does it. Both predict a valley; distinguishing them is an active question. Either way, the radius valley is direct evidence that atmospheric escape sculpts planetary populations, exactly as it sculpted Mars.

Hot Jupiters and packed systems

Hot Jupiters caused the first crisis. A Jupiter-mass planet cannot form at 0.05 AU: it is inside the frost line, the disk there cannot supply the solids or the gas, and the temperature is far too high. So it must have moved, which is how migration entered the standard picture, as Module 1 described.

How it moved is still argued. Disk migration has the planet spiralling inward while gas is still present, which predicts orbits that are circular and aligned with the star's spin. High-eccentricity migration has the planet flung inward by gravitational interaction with another planet or a companion star, then circularised by tides, which predicts a range of eccentricities and misalignments. Measurements of the angle between planetary orbits and stellar spin find plenty of misaligned and even retrograde hot Jupiters, so both routes probably operate.

A second hot Jupiter puzzle: many have radii substantially larger than models predict for their mass. The inflated radius problem has candidate explanations involving deposition of stellar energy deep in the atmosphere or ohmic heating from winds moving through the planet's magnetic field, and it is not settled.

The other structural surprise is compact multiple systems. TRAPPIST-1, an M dwarf 40 light years away, has seven roughly Earth-sized planets, all orbiting closer than Mercury does to the Sun, with the outermost completing an orbit in under 19 days. Kepler-11 has six planets inside the orbit of Venus. And such systems show a pattern nicknamed peas in a pod: planets within one system tend to be similar in size and regularly spaced, far more so than a random draw would produce. That regularity is a genuine constraint on formation models and it is not something our own system displays.

Key idea: The radius valley near 1.5 to 2 Earth radii is direct evidence that atmospheric escape sorts small planets into stripped rocky cores and envelope-retaining mini-Neptunes.

Is our solar system unusual?

This is where discipline is required. It is true that no exact solar system analogue has been found. It is not true that this constitutes evidence of rarity, and the reason is entirely about detection.

Consider what finding our own system from outside would require. Jupiter has a period of 11.86 years, so confirming it by radial velocity needs more than a decade of continuous, stable, high-precision monitoring; only a handful of programmes have baselines that long. Earth produces a wobble of 8.9 centimetres per second, below the current practical floor set by stellar noise, and its transit probability is 0.47 per cent with a depth of 84 parts per million occurring once a year. A survey like Kepler would have needed to observe our Sun for many years and would still, on the geometry alone, have missed Earth about 99.5 per cent of the time.

So the absence of solar system analogues in the catalogue is close to exactly what you would predict even if such systems were common. The correct statement is that we do not yet know whether our arrangement is typical, and answering it needs longer baselines and better precision, which is one of the things ESA's PLATO mission and next-generation spectrographs are for.

What we can say is that small planets are common. Occurrence-rate analyses of the Kepler sample suggest most stars have at least one planet, and that planets between Earth and Neptune in size are abundant. The specific quantity people care about, eta-Earth, the fraction of Sun-like stars with an Earth-sized planet in the habitable zone, is estimated somewhere in the range of roughly 10 to 60 per cent depending on the analysis. That is a wide range because it requires extrapolating past the edge of Kepler's sensitivity, and you should quote it as an order-of-magnitude statement rather than a number.

Reading an atmosphere

Detecting a planet is one thing; learning what its air is made of is another. The main technique is transmission spectroscopy. During a transit, a thin annulus of starlight passes through the planet's atmosphere on its way to us. At wavelengths where an atmospheric molecule absorbs, the atmosphere is effectively opaque and the planet looks slightly bigger, so the transit is slightly deeper. Measure transit depth as a function of wavelength and the variations map out the absorbing species.

Two related methods add information. In secondary eclipse, the planet passes behind the star, and subtracting the star-only measurement from the combined one isolates the planet's own thermal emission. Phase curves track brightness through a full orbit, revealing how heat is transported from dayside to nightside.

Worked example: how big is the signal?

The size of a transmission signal depends on how puffy the atmosphere is, measured by the scale height, the vertical distance over which pressure falls by a factor of e:

H = k T / (mu x m_H x g)

where k is Boltzmann's constant 1.381 x 10^-23, T is temperature, mu is mean molecular weight in atomic mass units, m_H is 1.674 x 10^-27 kg, and g is surface gravity.

A hot Jupiter. T = 1,200 K, mu = 2.3 for hydrogen and helium, g = 10 m/s^2.

Numerator: 1.381 x 10^-23 x 1,200 = 1.657 x 10^-20.

Denominator: 2.3 x 1.674 x 10^-27 x 10 = 3.85 x 10^-26.

H = 4.3 x 10^5 m, which is 430 kilometres.

The extra transit depth from the atmospheric annulus is approximately 2 x R_planet x H divided by R_star squared. For a planet of 1.27 Jupiter radii, 9.08 x 10^7 m, around a star of 0.9 solar radii, 6.26 x 10^8 m:

signal = (2 x 9.08 x 10^7 x 4.3 x 10^5) / (6.26 x 10^8)^2 = 7.82 x 10^13 / 3.92 x 10^17 = 2.0 x 10^-4.

That is 200 parts per million, comfortably within JWST's reach.

Now Earth. T = 255 K, mu = 29 for nitrogen and oxygen, g = 9.8 m/s^2.

Numerator: 1.381 x 10^-23 x 255 = 3.52 x 10^-21. Denominator: 29 x 1.674 x 10^-27 x 9.8 = 4.76 x 10^-25.

H = 7,400 m, about 7.4 kilometres, which is why Earth's atmosphere is a thin film.

signal = (2 x 6.371 x 10^6 x 7,400) / (6.96 x 10^8)^2 = 9.43 x 10^10 / 4.84 x 10^17 = 1.9 x 10^-7.

That is 0.2 parts per million, a thousand times smaller than the hot Jupiter signal. Look at what drove the difference: a hot, light, low-gravity atmosphere gives a huge scale height, while a cool, heavy, higher-gravity one gives a tiny one, and the small planet-to-star ratio squares the disadvantage. This one calculation tells you why hot Jupiters were characterised first, why M dwarf targets are chosen, and why an Earth twin around a Sun-like star is out of reach for JWST by roughly three orders of magnitude.

Key idea: Transmission signals scale as planet radius times atmospheric scale height over stellar radius squared, giving about 200 parts per million for a hot Jupiter and about 0.2 for Earth, which is the quantitative reason Earth analogues cannot be characterised today.

What JWST has actually done

JWST has transformed exoplanet atmospheres within its real limits. On the hot giant WASP-39b it delivered the first unambiguous detection of carbon dioxide in an exoplanet atmosphere in 2022, followed by sulfur dioxide, which is a photochemical product, meaning JWST is now observing chemistry driven by starlight in another planet's air. Full spectra of hot giants routinely show water, carbon monoxide, and clouds.

For rocky planets the news is more sobering. Observations of TRAPPIST-1b and TRAPPIST-1c measured dayside thermal emission consistent with little or no substantial atmosphere, suggesting these innermost planets were stripped, plausibly by the star's early activity. That is a real result and a cautionary one for M dwarf habitability.

Then there is K2-18b, which is the current live controversy and belongs here as a case study. It is a planet of about 2.6 Earth radii and 8.6 Earth masses in the habitable zone of an M dwarf. In 2023 a team led by Nikku Madhusudhan reported methane and carbon dioxide in its atmosphere and, more cautiously, a possible signal of dimethyl sulfide, a molecule produced on Earth almost exclusively by marine life. In 2025 the same group reported a stronger apparent signal of dimethyl sulfide or a related compound at around three sigma.

The response has been substantial and sceptical. Independent reanalyses of the same data by other groups have found the spectra fully consistent with no dimethyl sulfide at all, with the apparent signal sensitive to choices in data reduction and model fitting. A three sigma result is a roughly one in three hundred chance of arising by noise, which for an ordinary measurement is interesting and for a claim of this magnitude is nowhere near sufficient, particularly when many possible molecules are being searched over. Separately, the interpretation of K2-18b as a hycean world, with a hydrogen atmosphere over a liquid water ocean, is itself contested; other models fit the data with a hot mini-Neptune having a magma ocean and no habitable surface at all.

The field's position as of 2026 is that no biosignature has been detected on any exoplanet. K2-18b is an interesting target with a disputed spectral feature and a disputed structural interpretation. Note how precisely this rhymes with ALH84001 and with the Venusian phosphine: an exciting signal, a biological interpretation offered early, and a slow grinding process of independent checks. That process is the reason to trust the field, not a reason to distrust it.

Common misconceptions

"Super-Earths are just bigger Earths." The name describes radius. Many in the range are mini-Neptunes with thick hydrogen envelopes and no surface in any useful sense, which is why the mass measurement matters.

"No solar system analogues have been found, so ours is rare." Our methods are least sensitive to exactly that configuration. The non-detection is close to what you would expect even if such systems were common.

"JWST can look for life on Earth-like planets." An Earth twin around a Sun-like star produces a transmission signal near 0.2 parts per million, roughly a thousand times below what JWST measures on hot giants.

"Dimethyl sulfide was found on K2-18b." A tentative feature was reported at about three sigma and independent reanalyses find the data consistent with no such molecule. It is not an established detection.

Recap

Super-Earths and mini-Neptunes are the most common planets found and have no solar system counterpart, and the radius valley between about 1.5 and 2.0 Earth radii shows atmospheric escape sorting them into stripped cores and envelope-keepers by photoevaporation or core-powered mass loss. Hot Jupiters, at roughly 0.5 to 1 per cent of Sun-like stars, forced migration into formation theory, and their orbital misalignments suggest more than one migration route. Compact multiples such as TRAPPIST-1 show peas-in-a-pod regularity. Whether our own architecture is unusual is unresolved, because a Jupiter analogue needs a decade of radial velocity baseline and an Earth analogue needs 8.9 centimetres per second precision and lucky geometry. Atmospheric signals scale with the scale height H equal to kT over mu m_H g, giving 430 kilometres and about 200 parts per million for a hot Jupiter against 7.4 kilometres and 0.2 parts per million for Earth. JWST has detected carbon dioxide and photochemical sulfur dioxide on hot giants and found little atmosphere on the inner TRAPPIST-1 planets, and the disputed K2-18b dimethyl sulfide claim remains unestablished.

Sources

1. Fulton, B. J., Petigura, E. A., Howard, A. W., et al. (2017). The California-Kepler Survey III: A gap in the radius distribution of small planets. The Astronomical Journal, 154(3), 109. doi.org
2. JWST Transiting Exoplanet Community Early Release Science Team. (2023). Identification of carbon dioxide in an exoplanet atmosphere. Nature, 614, 649-652. doi.org
3. Madhusudhan, N., Sarkar, S., Constantinou, S., et al. (2023). Carbon-bearing molecules in a possible hycean atmosphere. The Astrophysical Journal Letters, 956(1), L13. doi.org
4. NASA Science. (n.d.). Exoplanet types and the search for another Earth. National Aeronautics and Space Administration. science.nasa.gov
5. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). Exoplanets everywhere: What we are learning. In Astronomy 2e (Section 21.6). OpenStax, Rice University. openstax.org

The big picture

Everything in this course has been building toward one question. Suppose a telescope returns a spectrum from a planet forty light years away. What feature in that spectrum would justify announcing that we had found life?

This is hard for a reason that runs through the whole course. We cannot go there, culture a sample, run a control, or repeat the experiment. We will have a handful of photons, a model of what a lifeless version of that planet would look like, and the standing obligation from Lesson 1 to rule out abiotic explanations before invoking biology. ALH84001, Venusian phosphine, and the K2-18b dispute have each shown what happens when that obligation is met slowly and publicly.

This last lesson covers what we would look for, what we have looked for, how to think about how much life there might be, and what obligations come with looking. It ends where the course began, with the discipline of separating evidence, inference, and speculation.

Key idea: A biosignature claim is only as strong as the abiotic alternatives that have been excluded, which means detecting life remotely is a modelling problem as much as an observational one.

What makes a good biosignature

A biosignature is an observable feature whose presence is best explained by life. To be useful it needs four things: it must be produced by life, detectable at interstellar distance, difficult to produce without life, and, most demandingly, we must be able to show that the abiotic routes are inadequate for that particular planet.

The obvious candidate is oxygen. Earth's atmosphere is 21 per cent molecular oxygen, essentially all of it produced by photosynthesis. Oxygen is chemically aggressive, reacting with rock and volcanic gases and almost anything else, so without continuous resupply it would vanish in geologically short order. A planet holding abundant oxygen is one where something is making it fast.

And oxygen is not sufficient, which is one of the most important results of the last two decades of astrobiology. Several abiotic routes can build up oxygen without any biology.

• On a planet undergoing a runaway greenhouse, water vapour reaches the upper atmosphere, ultraviolet light splits it, hydrogen escapes, and the leftover oxygen accumulates. This is early Venus, and models suggest it could produce hundreds of bars of abiotic oxygen.
• Around M dwarfs, whose ultraviolet spectra differ sharply from the Sun's, photolysis of carbon dioxide can generate substantial oxygen while the carbon monoxide by-product fails to recombine efficiently.
• A dry planet with no surface sink retains oxygen that would otherwise be scrubbed away.

So oxygen alone would be interesting and would not settle anything. The field's response has been to look for chemical disequilibrium instead: combinations of gases that react with each other and therefore cannot coexist for long without something continuously replenishing both. Oxygen together with methane is the classic pair, because in an oxygen atmosphere methane is destroyed within a decade or so. Finding both at appreciable levels means both are being made now, which is far harder to arrange abiotically than either alone. Earth shows exactly this signature, and it has for billions of years.

Weaker approaches exist. The red edge, the sharp rise in reflectance of chlorophyll-bearing vegetation just beyond the visible, is a real but very small feature of Earth's spectrum. Seasonal variation in gas abundances would be a temporal biosignature, and Earth's carbon dioxide does this as northern forests leaf out.

There are also antibiosignatures. Abundant carbon monoxide is the standard example: it is an excellent energy source that any plausible biosphere would eat, so finding a lot of it suggests nothing is.

Technosignatures

A technosignature is evidence of technology rather than of biology, and the trade is a good one to understand. Technosignatures are rarer, because they require not just life but technology, and they are far less ambiguous, because nature does not build radio transmitters.

The candidates:

• Radio emission, especially narrowband signals. Natural sources are broadband; a signal confined to a few hertz is not something known physics produces.
• Optical or infrared laser pulses, which could briefly outshine a star in a narrow band.
• Industrial atmospheric pollutants. Chlorofluorocarbons are the cleanest example in principle: no known abiotic source, strong infrared features, close to unambiguous. They would also be extraordinarily hard to detect at realistic abundances.
• Waste heat and megastructures, such as infrared excess from a star partially enclosed by collectors. Searches have found nothing anomalous that survived scrutiny.

State the comparison plainly: a biosignature is more likely to exist and harder to interpret, a technosignature less likely to exist and easier to interpret. A serious search does both.

Key idea: Oxygen has multiple abiotic production routes, so the strongest remote biosignature is chemical disequilibrium, such as oxygen and methane coexisting, since sustaining two mutually destructive gases requires continuous production of both.

SETI: sixty-five years of listening

The modern search began with a 1959 paper by Giuseppe Cocconi and Philip Morrison arguing that radio was the obvious interstellar channel and the 1420 megahertz hydrogen line an obvious frequency for strangers to agree on. In 1960 Frank Drake pointed an 85-foot dish at Tau Ceti and Epsilon Eridani for Project Ozma and heard nothing.

Since then: the Ohio State survey recorded the Wow! signal in August 1977, a strong narrowband burst lasting 72 seconds that matched expectations for an artificial source and has never been seen again, leaving it an unexplained one-off rather than a detection. Project Phoenix surveyed about a thousand nearby stars in the 1990s. Breakthrough Listen, funded with 100 million dollars from 2015, has run by far the most systematic search ever across radio and optical, with all data public. In 2020 a candidate called BLC-1, apparently from Proxima Centauri, was traced to human radio interference.

The result: no confirmed detection has ever been made.

How much does that tell us? Less than it feels like. The parameter space spans every direction on the sky, every frequency, every polarisation, every moment, and every sensitivity threshold. A 2018 analysis by Jason Wright and colleagues estimated that all SETI searches to date have examined roughly the equivalent of a hot tub's worth of Earth's oceans. That is a volume calculation, not a rhetorical trick, and it means the null result constrains very little.

The Drake equation, used properly

In 1961 Frank Drake wrote down an equation to organise the agenda for a small meeting. It has been misunderstood ever since.

N = R* x f_p x n_e x f_l x f_i x f_c x L

where N is the number of currently communicating civilisations in the galaxy, R* is the rate of star formation, f_p is the fraction of stars with planets, n_e is the average number of habitable planets per such system, f_l is the fraction of those where life appears, f_i is the fraction of those where intelligence appears, f_c is the fraction of those that communicate detectably, and L is how long they do so.

Notice the structure. The terms progress from astronomy to biology to sociology, and our knowledge collapses in exactly that order.

Term	Status in 2026
R*, star formation rate	Reasonably well measured, a few stars per year
f_p, fraction with planets	Now measured and close to 1; this is the genuine progress since 1961
n_e, habitable planets per system	Roughly constrained, order 0.1 to 1, depending on what habitable means
f_l, fraction where life arises	Completely unknown, for the reasons in Module 5
f_i, fraction reaching intelligence	Completely unknown
f_c, fraction that communicate	Completely unknown
L, communicating lifetime	Completely unknown

Worked example: why the equation cannot predict

Take a defensible optimistic set: R* = 2, f_p = 1, n_e = 0.2, f_l = 1, f_i = 0.1, f_c = 0.5, L = 10,000 years.

N = 2 x 1 x 0.2 x 1 x 0.1 x 0.5 x 10,000 = 200.

Two hundred communicating civilisations in the galaxy right now. Now take a defensible pessimistic set, keeping the astronomy identical and changing only the unknown terms: f_l = 10^-10, f_i = 10^-3, L = 100 years.

N = 2 x 1 x 0.2 x 10^-10 x 10^-3 x 0.5 x 100 = 2 x 10^-12.

Two in a trillion, which is to say we are alone in the galaxy and probably in a large volume around it. The two answers differ by fourteen orders of magnitude, and every input used was defensible, because the unknown terms are genuinely unknown rather than merely uncertain.

The conclusion is not that the equation is worthless. It is a bookkeeping device for our ignorance, and its value is diagnostic: it shows which measurement would narrow the answer most. That measurement is f_l, whether life arises given habitable conditions, which a second independent origin anywhere in our own solar system would transform. Find microbes on Enceladus unrelated to us, and f_l jumps from unknown to probably common in one result. That is why ocean worlds matter, and Drake intended the equation to structure a conversation, not to produce a number.

Key idea: Defensible inputs to the Drake equation span fourteen orders of magnitude in the answer, so it functions as a map of what we do not know rather than as a prediction, and the term that matters most is the probability that life arises at all.

The Fermi paradox

Over lunch at Los Alamos in 1950, Enrico Fermi is said to have asked, "Where is everybody?" The argument is about timescales. The galaxy is roughly 13 billion years old and about 100,000 light years across, so a civilisation expanding at even one per cent of light speed could cross it in a few tens of millions of years, a small fraction of galactic history. If technological civilisations were at all common, the galaxy should show signs of them. It does not.

The proposed resolutions sort into four families, and notice how testable each is.

1. They are not there. Some step is fantastically improbable: the origin of life, the eukaryotic cell, complex animals, or technology. This is the great filter, and it carries an uncomfortable corollary. If the filter is behind us we are extraordinarily lucky; if ahead, it is something that reliably ends technological civilisations. Indirectly testable, since an independent origin of life nearby would rule out the earliest filters.
2. They are there but do not expand or transmit. Interstellar travel may be prohibitively expensive, expansion may not be a stable goal, or civilisations may go quiet. Partly testable through better technosignature searches.
3. They are there and we have not noticed. The haystack argument above, or the zoo hypothesis in which we are deliberately left alone. The haystack version is testable; the zoo version is not, and unfalsifiable proposals should be labelled as such.
4. We are early. Small red dwarfs will burn for trillions of years, so most habitable planet-time lies in the future. On this view there is no paradox, just an early arrival.

Nobody knows which is right. Be suspicious of confidence here.

Planetary protection and the ethics of looking

Suppose there is life on Mars or in Europa's ocean. Then our spacecraft are a hazard to it, and it is potentially a hazard to us. Planetary protection is the international policy framework addressing this, developed under COSPAR, the Committee on Space Research, and grounded in Article IX of the 1967 Outer Space Treaty, which obliges parties to avoid harmful contamination of celestial bodies.

The policy sorts missions into categories by destination and mission type, with requirements rising steeply for bodies of astrobiological interest. Forward contamination, carrying terrestrial organisms outward, is the main concern; back contamination, returning something harmful, drives the containment requirements for sample return.

The practical consequences are substantial and expensive. The Viking landers were baked whole at 112 degrees Celsius for 30 hours, a level of sterilisation no mission since has matched, partly because it constrains what instruments you can fly. Galileo was deliberately destroyed in Jupiter's atmosphere in 2003 rather than left to drift and possibly hit Europa, and Cassini was destroyed in Saturn in 2017 to protect Enceladus and Titan. Curiosity and Perseverance are barred from the special regions where liquid water might transiently exist, because they are not clean enough. The tension is genuine: the places most worth investigating are the ones we are least permitted to touch.

Three further ethical questions are live, and this course will not resolve them for you.

Contaminating a second genesis. If Mars hosts an independent origin of life, it is the most scientifically valuable object humans have ever encountered, and it would be compromised by colonisation, or possibly just by a crewed mission, which cannot be sterilised because the crew is not. Crewed Mars exploration and planetary protection are in direct and unresolved conflict.

Resources and ownership. The Outer Space Treaty forbids national appropriation of celestial bodies but says little about extracted resources, and national laws passed since have asserted rights to what is mined. Enforcement is weak and the commercial sector is growing faster than the framework governing it.

Should we transmit? Deliberately broadcasting, sometimes called METI or active SETI, has been done a handful of times. Critics including Stephen Hawking argued that announcing ourselves to an unknown recipient is an asymmetric risk taken on behalf of everyone. Defenders reply that our broadcasts already leak and the added risk is small. The strongest point on either side is procedural: there is no mechanism by which humanity consents.

Common misconceptions

"Oxygen in an exoplanet atmosphere would prove life." Several abiotic routes produce oxygen, including ocean loss during a runaway greenhouse and carbon dioxide photolysis around M dwarfs. Disequilibrium pairs such as oxygen with methane are far stronger.

"Decades of SETI silence show nobody is out there." Searches to date have covered a minuscule fraction of the frequency, sky, time, and sensitivity space, comparable to sampling a hot tub of the oceans. The null result constrains very little.

"The Drake equation estimates how many civilisations exist." Its last four terms are unknown, and defensible inputs give answers spanning fourteen orders of magnitude. It organises ignorance rather than resolving it.

"The Wow! signal was a message." It was a strong unexplained narrowband burst lasting 72 seconds that has never recurred despite many searches, and unrepeatable single events cannot be confirmed.

Recap

A biosignature must be produced by life, detectable remotely, hard to produce abiotically, and have its abiotic alternatives actually excluded for the planet in question. Oxygen fails that last test alone, since runaway greenhouse ocean loss and M dwarf photochemistry generate it, so disequilibrium pairs such as oxygen with methane are stronger. Technosignatures are rarer but far less ambiguous. SETI has run since 1960 with no confirmed detection, though the searched fraction of parameter space is tiny. The Drake equation runs from well-measured astronomy to unknown biology and sociology, and defensible inputs yield answers from 200 down to two in a trillion, a spread of fourteen orders of magnitude, making it a map of ignorance whose most valuable missing term is the probability that life arises. The Fermi paradox has testable and untestable resolutions. Planetary protection, grounded in the 1967 Outer Space Treaty, has already cost us Galileo and Cassini deliberately, bars our rovers from the most interesting Martian ground, and conflicts unresolvedly with crewed exploration.

Sources

1. Fraknoi, A., Morrison, D., & Wolff, S. C. (2022). The search for extraterrestrial intelligence. In Astronomy 2e (Section 30.4). OpenStax, Rice University. openstax.org
2. Schwieterman, E. W., Kiang, N. Y., Parenteau, M. N., et al. (2018). Exoplanet biosignatures: A review of remotely detectable signs of life. Astrobiology, 18(6), 663-708. doi.org
3. NASA Astrobiology. (n.d.). Biosignatures and the search for life. National Aeronautics and Space Administration. astrobiology.nasa.gov
4. NASA Office of Planetary Protection. (n.d.). Planetary protection. National Aeronautics and Space Administration. sma.nasa.gov
5. United Nations Office for Outer Space Affairs. (1967). Treaty on principles governing the activities of states in the exploration and use of outer space. unoosa.org
6. Encyclopaedia Britannica. (n.d.). Search for extraterrestrial intelligence. britannica.com