The heat of fusion of ice is about 334 joules per gram, or equivalently around 80 calories per gram and 6.01 kilojoules per mole. That number represents the energy you need to pump into ice at 0 °C to turn it into liquid water at 0 °C, with no rise in temperature along the way. Modern calorimetry measurements have confirmed this value to within fractions of a percent, and it has remained one of the most reliably pinned-down constants in thermodynamics for well over a century. But the number itself only starts to get interesting once you see how large it really is relative to other thermal processes and how many fields depend on it.
What This Number Actually Tells You
Heat of fusion is the amount of energy absorbed when a solid turns into a liquid at its melting point, with the temperature held steady throughout. For ice, 334 J/g means that melting one gram of ice soaks up the same amount of energy it would take to heat that same gram of liquid water from 0 °C to about 80 °C. Think about that for a moment: the energy cost of merely breaking the solid apart, without warming anything, is almost as large as the energy cost of heating the resulting liquid from freezing to near boiling. That makes ice an unusually powerful absorber of heat compared with many other common solids.
The process works in reverse too. When liquid water freezes at 0 °C, it releases 334 J/g back into its surroundings. That released energy is why frost on a citrus orchard can actually slow overnight cooling: as a thin layer of water sprayed on the trees begins to freeze, it dumps heat into the fruit and the surrounding air, buffering the temperature. The same principle underlies a surprising range of practical applications, from keeping drinks cold to storing energy in buildings.
Why Ice Needs So Much Energy to Melt
Water molecules in ice are locked into an open, hexagonal crystal lattice held together by hydrogen bonds. Each molecule forms up to four hydrogen bonds with its neighbors, creating a rigid three-dimensional scaffold. Melting does not destroy all of those bonds at once, but it disrupts enough of them to collapse the orderly lattice into the disordered, denser arrangement of liquid water. Breaking a large fraction of those hydrogen bonds is energetically expensive, and that expense is reflected in the unusually high heat of fusion.
For comparison, most small organic molecules and many metals have heats of fusion well below water’s on a per-gram basis. Ethanol’s heat of fusion, for instance, is less than a third of water’s. The difference comes down to how strongly the molecules hold on to one another in the solid state. Water’s extensive hydrogen-bonding network is among the strongest intermolecular bonding arrangements found in common substances, which is why ice demands such a large energy input before it gives way.
How the Value Is Confirmed in Modern Labs
The 334 J/g figure was originally established through classical calorimetry experiments in the 19th century, but it continues to be tested and verified with increasingly precise instruments. The standard modern tool is a differential scanning calorimeter, or DSC, which measures tiny differences in heat flow between a sample and a reference as temperature or pressure changes at a controlled rate.
A 2022 study using a micro-DSC found the specific enthalpy of the ice-to-water transition to be within ±0.51 % of the accepted 333 J/g literature value, validating the experimental setup to a high degree of accuracy.1International Journal of Thermofluids. Calorimetry based temperature and specific enthalpy measurements associated with ice-water phase change in saline systems for freeze desalination Separate work using high-pressure DSC tested the ice-melting latent heat under elevated pressures and found the relationship between measured and reference values yielded a mean error of less than 2 %.2Journal of Food Process Engineering. HIGH‐PRESSURE DIFFERENTIAL SCANNING CALORIMETRY (DSC): EQUIPMENT AND TECHNIQUE VALIDATION USING WATER–ICE PHASE‐TRANSITION DATA The consistency of these results across different labs, instruments, and pressure conditions is part of what makes water’s phase-change data a go-to benchmark for calibrating new calorimeters in the first place.
How Salt and Other Dissolved Substances Change Things
Pure water freezes and melts at 0 °C with a clean, well-defined latent heat. Dissolve something in that water and the picture shifts. Saltwater, for example, does not undergo a single sharp phase transition. As the temperature drops, pure ice crystals begin to separate out from the brine, concentrating the remaining salt solution. The effective heat released or absorbed during this gradual process differs from the neat 334 J/g of pure water, because the ice forming out of a saline solution is still essentially pure H₂O but the liquid it leaves behind is increasingly salty, which changes its own thermal properties.
This distinction matters practically. In freeze desalination, engineers exploit the fact that ice rejects dissolved salts as it forms, producing relatively pure ice that can be melted for fresh water. Accurate measurements of how the enthalpy changes with salinity are critical for designing those systems. The micro-DSC study mentioned earlier was specifically aimed at mapping phase-change enthalpies across different salt concentrations, because using the pure-water value as a blanket assumption for saline systems introduces meaningful errors.1International Journal of Thermofluids. Calorimetry based temperature and specific enthalpy measurements associated with ice-water phase change in saline systems for freeze desalination
For everyday purposes, the dissolved substances most people encounter, like road salt or the minerals in tap water, lower the freezing point but do not dramatically change the amount of energy involved in the ice-to-water transition per gram of ice that actually forms. The main effect you notice is that the ice forms at a lower temperature and over a wider range, not that each gram of ice somehow stores dramatically more or less energy.
Supercooled Water Breaks the Simple Rule
Water can sometimes remain liquid below 0 °C without freezing, a state called supercooling. You have probably seen this if you have ever pulled a very still bottle of water from a freezer and watched it instantly crystallize when jostled. When supercooled water finally does freeze, you might expect it to release the same 334 J/g as water freezing at 0 °C. It does not.
Calorimetric measurements of water freezing at various degrees of supercooling show that the heat released during this nonequilibrium process can be considerably less than what you would predict from a standard reversible calculation.3PubMed. Heat of freezing for supercooled water: measurements at atmospheric pressure The reason is that a reversible phase transition assumes the system is at equilibrium throughout, gently nudging back and forth across the melting point. Supercooled freezing is a sudden, out-of-equilibrium event. The liquid and the forming ice are not at the same temperature at the instant of nucleation, and some of the energy budget goes into rapidly warming the nascent ice-water mixture back up toward 0 °C rather than being cleanly released as latent heat. The deeper the supercooling, the more pronounced this gap becomes.
This is not just a laboratory curiosity. In atmospheric science, cloud droplets routinely supercool to −20 °C or colder before ice crystals nucleate. The energy released when those droplets finally freeze affects local air temperature and convection inside clouds, influencing precipitation patterns. Getting the latent-heat accounting wrong in weather models, by simply plugging in 334 J/g regardless of supercooling, can introduce errors in forecasts of rain and snowfall intensity.
Ice as a Thermal Battery
Because 334 J/g is a large amount of energy per unit mass, ice has long been used as a way to store “coldness.” The modern version of this old idea shows up in commercial building cooling systems, where ice is made overnight using cheaper off-peak electricity and then melted during the day to help cool the building. The stored energy offsets the need to run large chillers during the hottest, most expensive hours.
Model-based control strategies for HVAC systems with ice-based cold thermal energy storage aim to reduce both energy costs and equipment sizes by shifting cooling loads away from peak demand periods.4Journal of Process Control. Energy efficient control of HVAC systems with ice cold thermal energy storage The appeal is straightforward: water is cheap, nontoxic, and its high heat of fusion means a relatively compact tank of ice stores a lot of cooling capacity. A cubic meter of ice at 0 °C absorbs roughly 306 megajoules as it melts, enough to keep a small office floor comfortable for an entire afternoon.
The concept scales up to seasonal storage as well. Research into seasonal ice thermal storage has explored forming ice during the heating season, when outdoor temperatures are below freezing, and then using that ice to supply free cooling in early summer as it melts.5Energy and Buildings. Optimal sizing and operation of seasonal ice thermal storage systems The ice is best formed during late winter, when low ambient temperatures maximize charging efficiency, and it can shave the seasonal cooling peak load enough to downsize the mechanical cooling equipment a building needs. The economics hinge on the heat of fusion: if ice stored less energy per kilogram, the tanks would need to be proportionally larger and the whole scheme would become less cost-effective.
Why Latent Heat Makes Cryopreservation Difficult
Preserving biological tissue by freezing it sounds conceptually simple, but the heat of fusion creates a persistent engineering headache. When you try to cool a whole organ, say a liver, below 0 °C, ice does not form everywhere at once. Crystals nucleate in some regions first, and as those crystals grow they release latent heat into the surrounding tissue. That locally released energy slows or even reverses the cooling rate in adjacent areas, making it extremely difficult to maintain a uniform temperature throughout the organ.
Research on cryopreservation of whole livers has identified the release of latent heat, along with complicated geometry and poor heat transfer through thick tissue, as a central obstacle to preserving organ viability after thawing.6PubMed. Cryopreservation of whole murine and porcine livers In a small cell suspension, the latent heat disperses quickly and the cooling rate stays relatively even. In a bulky organ, some regions freeze rapidly while others linger near 0 °C, bathed in the heat shed by nearby ice formation. The uneven freezing can rupture cells and damage fine blood-vessel networks, which is a major reason why whole-organ cryopreservation remains an unsolved problem despite decades of research. The same 334 J/g that makes ice great for cooling drinks makes it an adversary when you want controlled, gentle freezing of living tissue.
Pressure Changes the Numbers
At normal atmospheric pressure, ice melts at 0 °C and its heat of fusion is 334 J/g. Raise the pressure and both the melting point and the latent heat shift. Ice’s melting point actually drops slightly under pressure, one of water’s well-known anomalies, linked to the fact that ice is less dense than liquid water. The heat of fusion changes modestly with pressure too, though it remains in the same general range across the pressures encountered in food processing and geological systems.
The high-pressure DSC study tested ice melting at both atmospheric pressure (0.1 MPa) and 115 MPa and confirmed that the measured latent heat tracked reference literature values with high fidelity across that range.2Journal of Food Process Engineering. HIGH‐PRESSURE DIFFERENTIAL SCANNING CALORIMETRY (DSC): EQUIPMENT AND TECHNIQUE VALIDATION USING WATER–ICE PHASE‐TRANSITION DATA At extreme pressures, water can form different ice crystal structures (ice II, ice III, and so on), each with its own melting point and heat of fusion. These exotic ice forms exist mainly in the interiors of icy moons and in high-pressure laboratory experiments, not in your freezer, but they illustrate that the familiar 334 J/g belongs specifically to ordinary hexagonal ice (ice Ih) at or near atmospheric conditions.
Ice on Other Worlds
The heat of fusion matters well beyond Earth’s surface. Moons like Europa and Enceladus are thought to harbor liquid-water oceans beneath thick ice shells. In these environments, the exchange of latent heat at the ice-ocean boundary governs how the shell thickens or thins over time. Where warmer ocean water rises and contacts the underside of the ice, it delivers heat that melts ice from below. Where the water is cooler, ice can grow.
Numerical simulations of convection in subsurface oceans on icy moons show that the pattern of melting and freezing at the ice-ocean interface depends on latitude, because the moon’s rotation shapes how convective plumes organize. At polar regions, small mobile hotspots produce modest differences in ice thickness between melting and freezing zones, roughly 0.10 meters over 80 years. Near the equator, rotational effects stretch the convection into elongated bands, creating larger hotspots and more intense melting and freezing, with ice-thickness differences exceeding 0.24 meters over the same period.7Icarus. Convection in the subsurface ocean of icy moons and response of the upper ice layer All of these dynamics are ultimately driven by how much energy it takes to melt or form a given mass of ice. Without such a high heat of fusion, the ice shells on these moons would respond far more rapidly to small temperature fluctuations in the ocean below, and the geology of their surfaces would look very different.
Common Points of Confusion
A few misconceptions crop up regularly when people encounter the heat of fusion of ice for the first time. One is confusing it with the heat of vaporization of water, which is the energy needed to turn liquid water into steam. The heat of vaporization, at about 2,260 J/g, is roughly seven times larger than the heat of fusion. The two quantities measure different phase transitions and should not be mixed up, though they are both consequences of water’s strong hydrogen bonding.
Another common confusion is thinking that ice at −10 °C melts “faster” or absorbs “more energy” than ice at 0 °C. Ice below 0 °C does need extra energy to warm up to the melting point first, but the latent heat portion of the energy budget, the 334 J/g that does the actual melting, is the same regardless of where the ice started. A tray of ice cubes pulled from a −20 °C freezer absorbs a bit more total energy than one sitting right at 0 °C, but the extra energy goes to warming the solid ice, not to the phase change itself.
Finally, people sometimes assume that adding salt to ice lowers its heat of fusion. What salt actually does is depress the melting point, so the ice begins to melt at a lower temperature. The energy required to melt each gram of pure ice that forms in a salty mixture remains very close to 334 J/g. The confusion arises because the overall thermal behavior of a salt-ice mixture is more complex than pure ice melting, since you are simultaneously dissolving salt into the meltwater and shifting equilibrium concentrations. The per-gram latent heat of the ice crystals themselves, though, is not meaningfully altered by the presence of salt in the surrounding solution.