The Physics of Phase Transitions: Why Ice Melts, Water Boils, and Magnets Lose Their Power

The Most Ordinary Extraordinary Thing

You’ve watched water boil a thousand times. Kettle on, wait, bubbles form, steam rises, tea gets made. It’s so mundane that it barely registers as physics.

But here’s what’s actually happening. Liquid water — a dense, incompressible fluid where molecules slide past each other in a chaotic dance of hydrogen bonds — is transforming into steam, a gas where those same molecules fly freely at hundreds of metres per second, bouncing off the walls of their container, occupying a volume 1,600 times larger than the liquid they came from. The chemical identity doesn’t change. Every molecule is still H₂O. But the collective behaviour — the way the molecules relate to each other — changes completely.

And during the transition itself, something deeply strange happens: the temperature stops rising. You can pump energy into boiling water all day, and it stays at exactly 100 °C (at sea level) until every last drop has become steam. The energy goes in, but the temperature doesn’t budge. Where does the energy go?

That question — simple, practical, profound — leads directly into one of the richest areas of physics. Phase transitions aren’t just about melting ice and boiling water. They’re about superconductors losing their resistance, magnets losing their magnetism, the early universe cooling into the forces and particles we see today, and the deepest questions about order, symmetry, and why matter organises itself the way it does.

Latent Heat: The Hidden Energy

The energy that disappears during a phase transition doesn’t actually disappear — it’s doing work. Specifically, it’s breaking or rearranging the bonds between molecules.

In a solid crystal, every molecule sits at a fixed position in a regular lattice, held in place by intermolecular forces. In ice, these are hydrogen bonds — each water molecule forming up to four bonds with its neighbours. To melt ice, you need to supply enough energy to break a significant fraction of these bonds, freeing the molecules to move past each other while staying in contact. This energy — the latent heat of fusion — is 334 joules per gram for water. It doesn’t raise the temperature because it doesn’t increase the average kinetic energy of the molecules. It increases their potential energy — the energy stored in the configuration of the system.

Boiling requires far more energy: 2,260 J/g. That’s nearly seven times the latent heat of fusion, and the reason is clear once you think about it. Melting loosens the rigid crystal structure but keeps molecules close together — most hydrogen bonds are merely rearranged, not eliminated. Boiling requires completely separating molecules from each other, breaking essentially all intermolecular bonds. You’re converting a dense liquid where every molecule constantly touches its neighbours into a dilute gas where molecules are, on average, ten molecular diameters apart.

This is why steam burns are so dangerous — far worse than hot water burns at the same temperature. When steam condenses on your skin, it releases 2,260 J/g of latent heat directly into your tissue. A gram of steam at 100 °C delivers about 2,680 joules to your skin (latent heat plus cooling from 100 °C). A gram of water at 100 °C delivers only about 420 joules (just cooling from 100 °C). Steam carries more than six times the thermal energy of the same mass of boiling water.

The concept of latent heat was discovered by Joseph Black in the 1760s. Before Black, people assumed that adding heat always raised temperature. Black showed that heat could be absorbed or released without any temperature change during a phase transition — the heat was “latent” (hidden). This was a crucial insight that made the development of steam engines possible: understanding how much energy steam actually carries is essential for engineering an efficient engine.

First-Order vs. Second-Order: Two Flavours of Transformation

Not all phase transitions are created equal. Physicists classify them into two fundamentally different types.

First-order transitions are the ones you know: melting, boiling, freezing, condensation. They involve latent heat (a discontinuous jump in entropy), a volume change, and coexistence of two phases at the transition temperature. When ice melts, you can have ice and water sitting next to each other at 0 °C — two distinct phases coexisting in equilibrium. The transition involves a sudden rearrangement of molecular structure.

First-order transitions can also exhibit phenomena like supercooling and superheating. Pure water can be cooled carefully below 0 °C without freezing — it becomes a metastable supercooled liquid. Disturb it (tap the container, drop in a dust particle) and it freezes explosively. Similarly, very pure water in a clean container can be heated slightly above 100 °C without boiling — until a nucleation event triggers sudden, violent boiling (this occasionally happens in microwaves with very clean cups, and it’s genuinely dangerous).

Second-order transitions (also called continuous transitions) are subtler and, in many ways, more interesting. There’s no latent heat, no coexistence of phases, and no abrupt structural rearrangement. Instead, a property called the order parameter changes continuously from a non-zero value to zero (or vice versa) at the transition point.

The classic example is the ferromagnetic transition. Below the Curie temperature (770 °C for iron), a ferromagnet has a spontaneous magnetisation — the magnetic moments of atoms align cooperatively, and the material is magnetic. As the temperature increases toward the Curie point, the magnetisation decreases smoothly. At exactly the Curie temperature, it reaches zero, and above it the material is paramagnetic — no spontaneous order. There’s no latent heat, no sudden jump, no coexistence of magnetic and non-magnetic regions. The order just fades away continuously.

Second-order transitions have peculiar properties near the critical point. Fluctuations become enormous — regions of the material flicker between ordered and disordered states at all length scales. The correlation length (how far order extends) diverges to infinity. The specific heat diverges. Properties follow power laws with universal critical exponents that depend not on the specific material but only on general features like dimensionality and symmetry. This universality — the fact that water near its critical point and a magnet near its Curie temperature obey the same mathematical laws — is one of the most surprising and beautiful results in theoretical physics.

The Phase Diagram: A Map of Matter

Every substance has a phase diagram — a map showing which phase is stable at each combination of temperature and pressure. For water, it’s one of the most studied diagrams in all of science.

At low temperatures and normal pressure, water is solid (ice). Increase the temperature past 0 °C, and it melts to liquid. Increase further past 100 °C, and it boils to gas. That’s the familiar sequence. But the phase diagram reveals much more.

The solid-liquid boundary slopes to the left for water (higher pressure lowers the melting point) — this is unusual and is a consequence of ice being less dense than water. For most substances, the solid-liquid line slopes to the right (higher pressure raises the melting point). Water’s anomalous behaviour is the reason ice floats and why life in frozen lakes survives winter.

The liquid-gas boundary extends upward to the right and terminates at the critical point — 374 °C and 218 atmospheres for water. Beyond this point, the distinction between liquid and gas vanishes. You can go from liquid to gas without crossing any phase boundary by taking a path through the supercritical region. This means “liquid” and “gas” are not fundamentally different phases — they’re two regions of the same phase, separated by a first-order line that simply ends.

The solid-gas boundary exists at low pressures — this is where sublimation occurs (solid directly to gas, bypassing liquid). On Mars, where atmospheric pressure is about 0.6% of Earth’s, water ice sublimes directly to vapour. The “triple point” — where the solid-liquid-gas boundaries meet — is where all three phases coexist simultaneously. For water, this occurs at 0.01 °C and 611.7 pascals (about 0.6% of atmospheric pressure). The triple point of water is so precisely reproducible that it’s used as a calibration point for temperature scales.

What I find beautiful about the phase diagram is how much physics it encodes in a simple two-dimensional plot. Every point represents a specific combination of temperature and pressure. Every line represents a phase transition. Every region represents a stable phase. And the critical point — where a line just stops — represents something philosophically interesting: the absence of a sharp distinction where we expected one.

Symmetry Breaking: The Deep Structure of Phase Transitions

There’s a concept that unifies essentially all phase transitions, from boiling water to the Higgs mechanism, and it’s called symmetry breaking.

Consider liquid water. It looks the same from every direction and from every position — it has continuous translational and rotational symmetry. Now cool it below 0 °C. The molecules arrange themselves into a crystal lattice with specific orientations and spacings. The continuous symmetry is broken — the crystal has a preferred structure, preferred directions, preferred positions. The higher symmetry of the liquid is replaced by the lower symmetry of the crystal.

The crystal could have formed in any orientation — there’s nothing special about the direction it chose. But it had to choose one. The symmetry of the underlying physics (water molecules interact the same way in all directions) is higher than the symmetry of the ground state (the crystal has specific axes). This is spontaneous symmetry breaking.

Lev Landau formalised this idea in the 1930s. He showed that phase transitions can be understood as changes in symmetry, characterised by an order parameter — a quantity that is zero in the high-symmetry (disordered) phase and non-zero in the low-symmetry (ordered) phase. For melting/freezing, the order parameter is related to the density modulation of the crystal. For the ferromagnetic transition, it’s the magnetisation. For superconductivity, it’s the macroscopic quantum wave function of the superconducting condensate.

This framework is powerful because it’s universal. You don’t need to know the microscopic details of a system to understand its phase transition — you just need to know the symmetry of the order parameter and the dimensionality of the system. Different physical systems with the same symmetry properties exhibit the same critical behaviour, with the same critical exponents, the same scaling laws. A superfluid helium transition, a superconducting transition, and certain magnetic transitions all belong to the same universality class despite involving completely different physics at the atomic level.

The idea that symmetry breaking governs phase transitions turned out to have implications far beyond condensed matter physics. The Higgs mechanism — the process by which particles acquire mass in the Standard Model — is a phase transition in the vacuum itself, where the Higgs field acquires a non-zero value and breaks the electroweak symmetry. The early universe, cooling from the Big Bang, underwent a series of phase transitions as successive symmetries broke, giving rise to the distinct forces and particles we observe today.

From boiling water to the origin of mass. Same concept, radically different scale.

Exotic Phases: Beyond Solid, Liquid, Gas

The familiar trio of solid, liquid, and gas (plus plasma, the fourth state) barely scratches the surface of what matter can do.

Superfluids. Cool helium-4 below 2.17 kelvin (the lambda point), and it undergoes a phase transition to a superfluid state — a liquid with literally zero viscosity. A superfluid can flow through microscopic channels without any resistance, creep up the walls of a container and escape over the rim, and sustain persistent currents that flow forever without dissipation. The transition is a Bose-Einstein condensation: helium-4 atoms (which are bosons) crowd into the same quantum state, and the resulting macroscopic quantum wave function gives the fluid its frictionless flow. The lambda point gets its name from the shape of the specific heat curve near the transition — it looks like the Greek letter lambda (λ), with a sharp peak at the transition temperature.

Superconductors. Below a critical temperature, certain materials lose all electrical resistance. Current flows forever. Magnetic fields are expelled (the Meissner effect). The transition involves the formation of Cooper pairs — electrons that pair up through interactions with the crystal lattice and condense into a macroscopic quantum state. Like superfluidity, superconductivity is a phase transition characterised by a macroscopic quantum order parameter, and it breaks a specific symmetry (gauge symmetry). The critical temperatures range from millikelvins (for some exotic superconductors) to about 130 K (for high-temperature cuprate superconductors) — and the quest for room-temperature superconductors continues.

Bose-Einstein Condensates (BECs). Cool a dilute gas of bosonic atoms to nanokelvin temperatures (billionths of a degree above absolute zero), and the atoms condense into the lowest quantum energy state. The resulting BEC is a blob of matter where thousands of atoms behave as a single quantum entity — a superatom with a macroscopic wave function. BECs were first created experimentally in 1995 (earning Eric Cornell, Carl Wieman, and Wolfgang Ketterle the 2001 Nobel Prize). They’ve become essential tools for studying quantum phenomena at macroscopic scales, precision measurement, and quantum simulation.

Liquid crystals. Between the solid crystal phase and the isotropic liquid phase, some materials exhibit intermediate phases — liquid crystals — where molecules can flow like a liquid but maintain orientational order like a crystal. In the nematic phase, rod-shaped molecules all point roughly the same direction but have no positional order. In the smectic phase, they also arrange in layers. These phases are the basis of LCD technology: applying an electric field rotates the molecular orientation, changing how light passes through. Every phone screen, laptop display, and flat-screen TV is a device that manipulates a phase transition in a liquid crystal.

Quark-gluon plasma. Heat nuclear matter to about 2 × 10¹² kelvin (roughly 150,000 times the temperature at the centre of the Sun), and protons and neutrons dissolve. The quarks and gluons that were permanently confined inside hadrons become free, forming a quark-gluon plasma — a state of matter that existed for microseconds after the Big Bang. This phase has been recreated at CERN’s Large Hadron Collider and at Brookhaven’s RHIC. The transition from hadronic matter to quark-gluon plasma is believed to be a crossover (not a sharp phase transition) at zero baryon density, but may become a first-order transition at high baryon density — a prediction being tested by current and future particle accelerator experiments.

Nucleation: How Transitions Begin

Phase transitions don’t happen everywhere at once. They start at specific points — nucleation sites — and grow from there.

When water boils, bubbles don’t form spontaneously throughout the liquid. They nucleate at specific locations: scratches on the container wall, dissolved gas pockets, dust particles, surface imperfections. These sites provide a template where the new phase can form without requiring the enormous energy needed to create a bubble from scratch in the middle of a perfectly uniform liquid.

The physics is subtle. Creating a bubble of vapour inside a liquid requires creating a new interface (a surface), and that surface has an energy cost — the surface tension. For a tiny bubble, the surface-to-volume ratio is large, and the surface energy cost outweighs the thermodynamic benefit of the phase transition. Below a critical radius, bubbles spontaneously collapse. Above it, they grow. The critical radius depends on the degree of superheating: the more the liquid is above its boiling point, the smaller the critical radius, and the easier it is to nucleate bubbles.

This is the physics behind the Mentos-in-Diet-Coke geyser. The Mentos candy has a surface covered in microscopic pits — millions of nucleation sites per candy. When dropped into the supersaturated carbonated drink, these sites trigger explosive nucleation of CO₂ bubbles, and the resulting foam erupts spectacularly. The reaction isn’t chemical — it’s purely physical. It’s nucleation.

Cloud formation works the same way. Water vapour in the atmosphere needs nucleation sites — dust particles, sea salt aerosols, pollution particles — to condense into droplets. Without these condensation nuclei, air can become supersaturated with water vapour without forming clouds. This is the principle behind cloud seeding: introducing silver iodide particles into clouds provides additional nucleation sites, promoting precipitation.

What Phase Transitions Teach Us

Phase transitions sit at the crossroads of everyday experience and deep theoretical physics. Boiling a kettle and the origin of mass in the universe are governed by the same mathematical framework. That’s not a loose analogy — it’s a precise statement about symmetry, order parameters, and the way nature organises itself at the most fundamental level.

What I find most compelling is the concept of emergence. A single water molecule doesn’t have a phase. “Solid” and “liquid” and “gas” are not properties of individual molecules — they’re collective phenomena, properties that emerge only when vast numbers of molecules interact. The phase transition doesn’t happen to any individual molecule. It happens to the system as a whole.

This is a profound lesson about how complexity arises from simplicity. The interactions between water molecules are simple: hydrogen bonds, van der Waals forces, electrostatic attraction and repulsion. But the collective behaviour of 10²³ molecules interacting simultaneously produces phenomena — melting, boiling, critical opalescence, superfluidity — that no analysis of individual molecules could predict.

Thermodynamics gives us the macroscopic rules. Statistical mechanics gives us the microscopic foundation. And the theory of phase transitions — built by Landau, Onsager, Wilson, Kadanoff, Fisher, and others — gives us the bridge between scales, showing how microscopic interactions produce macroscopic transformations.

Every time you watch ice melt in a glass of water, you’re watching a phase transition. Molecules breaking free of their lattice. Order dissolving into disorder. Energy hiding in broken bonds. And if you look closely enough — with the right equations and the right perspective — you’re watching the same physics that shaped the universe in its first microseconds of existence.

All in a glass of water on a warm afternoon.