eduKate Learning Manual
Science | Edge Cases Science | Earth, Water & Atmosphere
Understand → Observe → Explain → Test → Transfer → Go Deeper

Supercooled Water

How Liquid Water Can Exist Below Its Freezing Point

Big Question: If ice is the stable phase below the equilibrium freezing temperature, why can water remain liquid—and what finally triggers freezing?

This Learning Manual begins with Primary states of matter and grows into Secondary particle models, crystallisation, nucleation theory, atmospheric microphysics and the engineering problem of aircraft icing.

Quick Answer

Freezing requires more than a low temperature. Liquid water must begin forming an ordered ice crystal. Tiny embryonic crystals are difficult to stabilise because creating a new ice–water interface costs energy. A suitable surface, impurity or pre-existing ice can lower the barrier by helping molecules organise into a viable nucleus.

If no effective nucleus forms, water can remain in a metastable liquid state below the equilibrium melting/freezing temperature. Once nucleation succeeds, ice growth can proceed rapidly and release latent heat.

Cold enough is necessary. A successful nucleus is the trigger.

What You Will Learn

• Why crossing 0 °C does not guarantee instantaneous freezing.
• What metastability means.
• Why crystals need nuclei.
• How homogeneous and heterogeneous nucleation differ.
• Why surfaces and impurities can promote freezing.
• Why freezing releases latent heat.
• How supercooled droplets exist in clouds.
• Why supercooled water is a serious aircraft-icing hazard.
• How ice shape can alter airflow over a wing.
• Why tiny droplets behave differently from a tray of ordinary freezer water.
• How scientists study supercooled water and icing.
• How the simple school model remains useful when its boundary is stated clearly.

Part 1 — The Freezing Point Is an Equilibrium Boundary

At roughly one atmosphere of pressure, pure water and ordinary ice have an equilibrium melting/freezing temperature close to 0 °C.

That means that at the boundary, solid and liquid can coexist under equilibrium conditions. Below it, ice is thermodynamically favoured over liquid water.

But thermodynamics does not automatically tell us how fast the transition will occur.

Thermodynamics asks: Which state is favoured?
Kinetics asks: How quickly can the system get there?

Supercooling lives in the gap between those questions.

Part 2 — Ice Is an Ordered Structure

Liquid water is not a random gas-like swarm. Its molecules continually form and break hydrogen-bonded arrangements. But those arrangements are dynamic and lack the long-range crystalline order of ordinary ice.

To freeze, a local cluster of water molecules must become organised in a way that can grow into a stable ice crystal.

That sounds simple. At microscopic scale, it creates an energetic problem.

Part 3 — Why the First Tiny Crystal Is Hard to Make

When a tiny ice embryo forms inside liquid water, two effects compete.

• Forming the solid phase becomes increasingly favourable as the water is cooled below equilibrium.
• Creating a new boundary between ice and liquid costs interfacial energy.

A very small embryo has lots of surface relative to its volume. The interfacial cost can dominate, so the cluster disappears.

If a cluster grows beyond a critical size, the energetic advantage of the solid interior can overcome the surface penalty. The nucleus becomes viable and crystal growth can continue.

The first successful crystal is harder than the next million layers.

Part 4 — Homogeneous Nucleation

Homogeneous nucleation means the critical ice nucleus forms within the liquid without relying on a foreign surface to organise it.

This requires substantial supercooling because the liquid must overcome the nucleation barrier on its own. In very clean small droplets, liquid water can persist far below 0 °C before spontaneous freezing becomes likely.

Do not memorize one absolute “lowest temperature” as though every droplet must behave identically. Freezing is probabilistic, depends on droplet size and timescale, and different experimental conditions produce distributions rather than one magical switch.

Part 5 — Heterogeneous Nucleation

In the everyday world, freezing often begins through heterogeneous nucleation. A foreign surface helps reduce the energetic cost of forming the new phase.

Possible nucleating surfaces include certain mineral dust particles, biological particles, container imperfections or existing ice.

Not every speck of dust is equally effective. Surface chemistry, crystal structure, roughness, size and temperature all influence ice-nucleating ability.

An impurity can matter not because it “makes water colder,” but because it provides a better place for ice to begin.

Part 6 — Metastable Does Not Mean Impossible

A metastable state can persist even though another state has lower free energy under the same conditions.

Imagine a ball sitting in a shallow valley separated by a hill from a deeper valley. The deeper valley is more stable, but the ball cannot reach it unless something carries it over the hill.

For supercooled water, the deeper valley represents ice. The hill represents the nucleation barrier.

This analogy is limited—the molecular system is statistical, not a literal rolling ball—but it captures the idea of an activation barrier.

Part 7 — What Happens When Freezing Finally Begins?

Once a viable nucleus forms, surrounding water molecules can join the growing crystal.

Freezing releases latent heat. That means the growing ice can warm the local system toward the equilibrium freezing temperature even while more liquid becomes solid.

This surprises learners because “freezing” sounds like a process that should only remove heat. In reality, heat must leave the system overall, but the phase transition itself releases energy as the more ordered solid structure forms.

supercooled liquid → nucleation → rapid crystal growth → latent heat released.

Part 8 — Why Clouds Can Contain Liquid Water Below 0 °C

Clouds contain tiny droplets, not miniature cups of water. Many droplets are micrometres across and can remain liquid at subfreezing temperatures if they lack an effective ice nucleus.

Clouds can therefore contain:

• liquid droplets above 0 °C;
• supercooled liquid droplets below 0 °C;
• ice crystals;
• mixtures of supercooled droplets and ice.

This mixed-phase behaviour matters for precipitation, cloud lifetime, radiation and climate.

NASA describes supercooled liquid-water clouds as a known aviation hazard and studies them using aircraft, remote sensing and icing facilities.

NASA GISS — Supercooled liquid-water clouds →

Part 9 — The Aircraft-Icing Problem

An aircraft flying through a cloud of supercooled droplets can provide exactly the surfaces those droplets need.

Droplets strike the leading edges of wings, tail surfaces, engine components, sensors or propellers. Depending on droplet size, temperature, surface condition and airflow, some or all of the water can freeze and build an ice layer.

That changes the aircraft’s geometry.

water microphysics → ice accretion → changed surface shape → changed airflow → changed lift and drag.

NASA’s Icing Research Tunnel can create controlled supercooled-water clouds at subfreezing temperatures to study icing physics and test protection systems.

NASA Glenn — Icing Research Tunnel →

Part 10 — A Thin Layer of Ice Can Matter More Than Its Mass Suggests

Aircraft icing is not dangerous simply because ice is heavy.

Even relatively small amounts of rough or strategically placed ice can disrupt the smooth airflow over an aerodynamic surface. That can increase drag, reduce lift and alter stall behaviour.

This gives a general engineering lesson:

Where material accumulates can matter as much as how much material accumulates.

Part 11 — Supercooling Is Not the Same as Freezing-Point Depression

Salt can lower water’s equilibrium freezing point. That is freezing-point depression.

Supercooling is different. In supercooling, the liquid remains below the relevant equilibrium freezing temperature because nucleation has not yet succeeded.

Phenomenon	What changed?
Freezing-point depression	Composition changes the equilibrium phase boundary.
Supercooling	The liquid persists metastably below the equilibrium boundary.

These can occur together in real systems, but they are conceptually different.

Part 12 — Supercooling Is Not the Same as “Cold Water That Forgot to Freeze”

It is tempting to personify the phenomenon: “the water doesn’t know it should freeze.”

That language is memorable but scientifically weak. Molecules do not know the temperature. The state persists because the microscopic pathway to a stable crystal has an activation barrier.

A better sentence is:

The liquid is metastable because a sufficiently large ice nucleus has not yet formed.

Part 13 — Why Droplet Size and Time Matter

Nucleation is statistical. A larger volume contains more molecules and more opportunities for a rare nucleation event or an active impurity. A longer observation time also gives more opportunities.

So a statement such as “water freezes spontaneously at exactly X °C” can hide an important probability distribution.

Scientists therefore report conditions carefully: droplet size, cooling rate, purity, pressure, observation time and nucleation probability.

Part 14 — Follow One Supercooled Droplet

1. A tiny cloud droplet forms from water vapour around an aerosol particle.
2. The surrounding air cools below 0 °C.
3. The droplet remains liquid because no effective ice nucleus has formed.
4. Molecules continue moving and rearranging within the liquid.
5. The droplet travels with the cloud.
6. It collides with an aircraft wing.
7. The solid surface helps trigger freezing.
8. An ice layer begins to accrete.
9. Later droplets strike the growing ice and add mass and roughness.
10. The changed surface alters local airflow.

A microscopic delay in crystallisation can therefore become a macroscopic engineering problem.

A Text Diagram You Can Draw Anywhere

temperature falls below 0 °C
          ↓
   liquid water droplet
          ↓
  no effective ice nucleus
          ↓
      SUPERCOOLED
          ↓
 collision / ice nucleus / suitable surface
          ↓
      NUCLEATION
          ↓
      ice crystal grows
          ↓
 latent heat released + liquid becomes solid

Boundary: this is a conceptual pathway. Real freezing probabilities depend on temperature, droplet size, composition, surfaces and time.

Think Like a Scientist: How Do We Study Something That Freezes Unpredictably?

• Controlled cooling measures when droplets freeze.
• Large numbers of droplets reveal statistical distributions rather than one anecdote.
• Microscopy observes ice growth.
• High-speed imaging captures rapid freezing events.
• Different particle types test ice-nucleating efficiency.
• Wind tunnels reproduce icing conditions on aerodynamic surfaces.
• Remote sensing identifies supercooled liquid-water clouds.
• Aircraft measurements sample cloud temperature, droplet size and liquid-water content.

Because nucleation is probabilistic, replication is essential.

Observation vs Inference

• Observation: a droplet is liquid at −10 °C.
• Observation: it freezes after contact with an ice crystal.
• Inference: the first droplet was supercooled and lacked a successful ice nucleus before contact.
• Further test: repeat with many droplets and different nucleating surfaces.

Common Misconceptions and Better Models

Misconception	Why it sounds plausible	Better model
Any water below 0 °C must be ice.	0 °C is taught as the freezing point.	Liquid water can persist metastably below equilibrium if nucleation is delayed.
Supercooling changes the chemical formula of water.	The behaviour is unusual.	It remains H₂O in the liquid phase.
Dust freezes water by making it colder.	Dust can trigger freezing.	Some particles provide surfaces that lower the nucleation barrier.
Equilibrium and kinetics are the same thing.	Both affect phase change.	Equilibrium describes favourability; kinetics describes pathway and rate.
Supercooling and salt lowering the freezing point are identical.	Both allow liquid below 0 °C.	One is metastability; the other shifts the equilibrium boundary.
Aircraft icing is mainly dangerous because ice adds weight.	Ice is heavy.	Surface roughness and shape changes can strongly disrupt aerodynamics.
Every impurity nucleates ice equally well.	Impurities can trigger freezing.	Ice-nucleating activity depends on material and surface properties.

Checkpoint Questions

1. What is supercooled water?
2. Why does ice formation require nucleation?
3. What is the interfacial-energy problem for a tiny ice embryo?
4. What is homogeneous nucleation?
5. What is heterogeneous nucleation?
6. What does metastable mean?
7. Why does freezing release latent heat?
8. Why can cloud droplets remain liquid below 0 °C?
9. How can supercooled droplets create aircraft icing?
10. Why is supercooling different from freezing-point depression?
11. Why do droplet size and observation time matter?
12. Why should nucleation experiments use many trials?

Apply It — Three Containers

• A: ordinary tap water in a scratched container.
• B: highly purified small droplets in a clean experimental chamber.
• C: water containing an efficient ice-nucleating particle.

All are cooled below 0 °C. Which system is most likely to remain liquid longest? What additional information would you want before predicting confidently?

Answer Key

Open after attempting the questions

1. Liquid water below its equilibrium freezing temperature.
2. A stable crystal requires an initial ordered nucleus large enough to grow.
3. Creating new ice–water interface costs energy and destabilises very small embryos.
4. Ice nucleation without a foreign nucleating surface.
5. Ice nucleation assisted by a foreign surface or particle.
6. A state that persists despite another state being thermodynamically more stable.
7. Formation of the solid phase releases latent heat.
8. Small droplets may lack effective ice nuclei.
9. They can freeze on impact with aircraft surfaces and build aerodynamically disruptive ice.
10. Supercooling delays transition below an existing boundary; dissolved substances can move the boundary itself.
11. Nucleation is probabilistic; volume and time change the number of opportunities.
12. Many trials reveal a distribution and reduce the risk of mistaking chance for a rule.

Application: B is generally the strongest candidate for deeper supercooling, but outcome depends on purity, droplet size, cooling rate, pressure and microscopic surface conditions.

Can You Explain WHY?

• Why can a lower-energy phase fail to appear immediately?
• Why does a surface sometimes trigger freezing?
• Why can supercooled water release heat while becoming ice?
• Why are tiny cloud droplets good places to find supercooling?
• Why can a small amount of ice seriously affect a wing?
• Why is a one-trial freezer demonstration weak evidence for a universal freezing temperature?

Singapore Connection

Singapore’s near-surface atmosphere is warm, so supercooled clouds are not an everyday ground-level experience. But Singapore is one of the world’s major aviation hubs. Aircraft departing, arriving and flying long routes encounter atmospheric conditions far beyond tropical street level.

The lesson therefore helps connect a warm equatorial city to cold cloud microphysics encountered in global aviation.

Primary Science Bridge

• water can change between solid and liquid;
• cooling can cause freezing;
• temperature can be measured;
• particles continue moving even when we cannot see them;
• one observation should not automatically become a universal rule.

The edge-case extension is: cooling below the usual freezing point can make freezing favourable without forcing it to begin instantly.

Secondary and JC Bridge

Core idea	Higher-resolution route
Freezing	Crystallisation and phase-transition kinetics
Particles	Hydrogen-bond networks and crystal order
Energy	Latent heat, free energy and activation barriers
Impurities	Heterogeneous nucleation and surface chemistry
Clouds	Atmospheric microphysics and mixed-phase clouds
Forces/airflow	Aerodynamic consequences of ice accretion

Deep Science Window — Critical Nucleus Size

Classical nucleation theory describes a competition between a favourable bulk free-energy change and an unfavourable surface-energy cost. The balance produces a critical nucleus size. Clusters smaller than this tend to disappear; clusters larger than it can grow.

As supercooling increases, the thermodynamic driving force for freezing becomes stronger and the nucleation barrier generally falls.

Deep Science Window — Ice Nuclei Can Be Biological

Some biological particles and proteins can nucleate ice efficiently. This connects microbiology, atmospheric science and precipitation. It is a reminder that disciplinary boundaries in textbooks do not exist in the world itself.

Deep Science Window — Mixed-Phase Clouds Matter to Climate

Clouds containing both ice and supercooled liquid water can reflect sunlight, emit infrared radiation and produce precipitation differently from all-liquid or all-ice clouds. Correctly representing their phase composition remains important in weather and climate models.

microscopic nucleation probability → cloud phase → cloud lifetime and radiation → planetary energy balance.

Evidence Boundaries

• Below 0 °C ≠ automatically solid. Metastable liquid can persist.
• Supercooling ≠ lowered equilibrium freezing point. These are different mechanisms.
• One impurity ≠ one universal nucleation temperature. Ice-nucleating ability varies.
• One freezing event ≠ deterministic law. Nucleation is statistical.
• Aircraft icing ≠ weight problem only. Surface shape and roughness matter.
• Metastable ≠ stable forever. A suitable fluctuation or surface can trigger transition.
• Simple 0 °C rule ≠ useless. It remains a valuable equilibrium reference when conditions are stated.

Manual Summary — KNOW → CONNECT → EXPLAIN → APPLY → CHECK

KNOW

Know supercooling, metastability, nucleus, crystallisation, heterogeneous nucleation, homogeneous nucleation and latent heat.

CONNECT

Connect temperature to phase stability, nucleation to kinetics, crystal growth to latent heat, and droplets to cloud and aviation systems.

EXPLAIN

Explain why a liquid can persist below its equilibrium freezing temperature without claiming the freezing point has disappeared.

APPLY

Predict how purity, surfaces, droplet size, temperature and time affect freezing probability.

CHECK

Ask whether you are discussing equilibrium, kinetics or both.

Research Sources and Further Reading

• NASA Glenn — Icing Research Tunnel
• NASA — Sensors Detect Icing Conditions to Help Protect Airplanes
• NASA GISS — Polarized view of supercooled liquid-water clouds

---

eduKate Learning Manuals are free educational material built so a simple school boundary can become the entrance to a deeper model rather than a fact to memorise and forget.