The specific heat of ice is approximately 2.09 joules per gram per degree Celsius (J/g·°C) at temperatures near 0 °C. That number means it takes about 2.09 joules of energy to raise one gram of ice by one degree. This is roughly half the specific heat of liquid water, which sits around 4.18 J/g·°C, and understanding why that gap exists and what it means in practice turns out to be more interesting than the number alone.
What That Number Actually Tells You
Specific heat is a measure of how stubbornly a substance resists changing temperature. A high specific heat means a material soaks up a lot of energy before it gets noticeably warmer; a low one means it heats up quickly. Ice, at 2.09 J/g·°C, falls in a middle range among common solid materials. It heats up faster than liquid water does, but it still absorbs a fair amount of energy per degree of temperature change compared to most metals or dry rock.
In everyday terms, if you pull a chunk of ice from a freezer set to −18 °C, that ice has to warm through 18 degrees before it even reaches the melting point. The energy needed to accomplish that warming is governed by the specific heat. For 100 grams of ice going from −18 °C to 0 °C, you would need roughly 3,760 joules just to bring it to the melting point, before any melting begins. That is a modest amount of energy compared to what comes next (actually melting the ice), but it is not trivial, and it matters in situations ranging from defrosting food to modeling glacial energy budgets.
How the Value Shifts With Temperature
The 2.09 figure is accurate near 0 °C, but ice does not have the same specific heat at every temperature. As ice gets colder, its specific heat drops. At around −40 °C it falls to roughly 1.7 J/g·°C, and by −80 °C it is lower still. Precise calorimetric measurements across the range from 0 °C down to −78.5 °C, achieved with accuracy better than ±0.05%, have confirmed this steady decline.1Canadian Journal of Research. SPECIFIC HEATS AND LATENT HEAT OF FUSION OF ICE The drop is smooth rather than abrupt, and it follows a well-understood pattern tied to how molecular vibrations behave at lower energies.
The practical upshot: if you are doing a calculation that involves very cold ice, say dry-ice-chilled samples in a lab or ice on a polar surface at −50 °C, using the 2.09 value will overestimate the energy required to warm that ice. For most kitchen and classroom purposes the temperature range is narrow enough that 2.09 works fine, but engineers and researchers working with deeply frozen materials use temperature-dependent curves rather than a single constant.
Why Ice Has a Lower Specific Heat Than Liquid Water
The difference between ice at 2.09 and liquid water at 4.18 is striking, given that both are just Hâ‚‚O. The explanation comes down to how the molecules are arranged and how freely they can move. In liquid water, molecules are in constant tumbling motion, forming and breaking hydrogen bonds on timescales of picoseconds. That chaotic environment creates many ways for incoming energy to be absorbed: molecules can translate, rotate, and vibrate, and the rapidly reshuffling hydrogen-bond network can soak up energy as bonds stretch and reform.
In ice, the molecules are locked into a rigid crystalline lattice. Each molecule sits in a fixed position connected to four neighbors through hydrogen bonds. There are fewer ways for energy to distribute itself because translational and rotational motion is largely frozen out. The energy goes mostly into vibrational modes of the lattice, and that is a smaller energy sink than the full suite of motions available in the liquid. Neutron-scattering studies of ice have revealed that the hydrogen-bond network is not perfectly uniform: evidence points to two distinct types of hydrogen bonds, with different strengths, distributed in roughly a 2-to-1 ratio of strong to weak bonds throughout the crystal.2Nature. Evidence for two kinds of hydrogen bond in ice This structural detail influences how vibrational energy propagates through the lattice and contributes to ice’s overall thermal behavior.
The Latent Heat Trap
One of the most common points of confusion when people look up the specific heat of ice is mixing it up with the latent heat of fusion. These are fundamentally different quantities. Specific heat describes energy per degree of temperature change. Latent heat of fusion describes the energy needed to melt ice at 0 °C into water at 0 °C, with no temperature change at all. The latent heat of fusion for ice is about 334 joules per gram, which is enormous compared to the specific heat.
To put this in perspective, warming one gram of ice from −1 °C to 0 °C takes about 2.09 joules. Then melting that same gram at 0 °C takes 334 joules, roughly 160 times as much energy, while the thermometer does not budge. This is why ice is so effective at cooling drinks: most of the cooling power comes from the melting process, not from the ice warming up beforehand. If you only accounted for specific heat and forgot the latent heat, you would drastically underestimate how much energy ice can absorb.
The same calorimetric measurements that refined the specific heat of ice across its temperature range also yielded high-precision values for the latent heat of fusion, because the two quantities are determined in the same experimental sweep.1Canadian Journal of Research. SPECIFIC HEATS AND LATENT HEAT OF FUSION OF ICE They are measured together but represent completely different physical processes, and confusing them leads to errors in everything from homework problems to industrial refrigeration design.
Units You Will Encounter
The specific heat of ice shows up in different unit systems depending on where you encounter it, and the number looks quite different in each one:
- J/g·°C: 2.09 (the most common form in modern science and engineering)
- J/kg·K: 2,090 (same value scaled up to kilograms; degrees Celsius and kelvins have the same size, so the number is identical aside from the factor of 1,000)
- cal/g·°C: 0.50 (since one calorie equals about 4.184 joules, the calorie-based value is roughly half the joule-based one)
- BTU/lb·°F: 0.50 (the British thermal unit system happens to give almost the same numerical value as calories per gram per Celsius degree, by coincidence of unit definitions)
- J/mol·K: 37.7 (per mole of water, useful in chemistry contexts where you are counting molecules rather than weighing samples)
The calorie-based value of 0.50 is easy to remember and is still widely used in food science and older reference tables. If you see the specific heat of ice quoted as “0.5” with no units, it is almost certainly in cal/g·°C.
How Food Engineers Use the Value
The specific heat of ice is not just a physics classroom staple. It plays a central role in food engineering, where predicting how quickly food freezes, how much energy a refrigeration system needs, and how temperature distributes through a frozen product all depend on knowing the thermal properties of the ice within the food. Frozen food is not pure ice; it is a complex mixture of ice crystals, unfrozen water, proteins, fats, and carbohydrates. But the ice fraction dominates the thermal behavior once a food is solidly frozen.
Reliable equations for the thermophysical properties of foods during cooling and freezing, including specific heat capacity, thermal conductivity, and enthalpy, have been developed based on generalized parameters like moisture content and freezing temperature.3Journal of Food Engineering. Predictive equations for thermophysical properties and enthalpy during cooling and freezing of food materials In these models, the specific heat of the ice component is one of the key inputs. Getting it wrong by even a small margin can cascade into significant errors in predicted freezing times, which in turn affects food safety, texture, and energy costs.
Consider a commercial blast freezer processing thousands of kilograms of chicken per hour. The engineers designing that system need to know how much energy the refrigeration unit must extract per minute. Part of that energy cools the chicken from its initial temperature to its freezing point, part goes into the phase change (latent heat), and part continues cooling the now-frozen product to its target storage temperature. That last stage, cooling already-frozen food, is where the specific heat of ice directly governs the calculation. Underestimate it and the freezer runs too briefly, leaving product cores warmer than intended. Overestimate it and the system is over-engineered, wasting capital and electricity.
How Ice Compares to Other Common Solids
Ice at 2.09 J/g·°C is a surprisingly good heat absorber for a solid. Most familiar solid materials have lower specific heats. Iron sits around 0.45 J/g·°C, copper at 0.39, aluminum at 0.90, and granite at roughly 0.79. Glass is about 0.84. Even wood, which varies by species and moisture content, typically ranges from 1.2 to 2.0 J/g·°C. Ice beats nearly all of them, which is part of why frozen ground and glacial ice play such large roles in regional and global energy balance: ice-covered landscapes store and release substantial amounts of thermal energy as their surface temperatures fluctuate.
The reason ice’s specific heat is high for a solid circles back to its hydrogen-bonded structure. Most solids are held together by metallic bonds, ionic bonds, or weaker van der Waals forces, all of which create vibrational landscapes that absorb less energy per gram per degree. Ice’s network of hydrogen bonds, while not as energy-absorbent as the chaotic liquid form, still provides a richer set of vibrational modes than you find in a metal lattice.
What Happens at the Boundaries
A few edge cases are worth knowing about, especially if you are doing calculations that involve extreme conditions rather than everyday ice.
Very near 0 °C, the specific heat of ice increases slightly as the crystal lattice approaches its melting point and pre-melting effects begin to loosen the surface layers. This is a subtle effect but measurable with precision calorimetry. At the other extreme, near absolute zero, the specific heat of ice drops toward zero in accordance with the general thermodynamic principle that all heat capacities vanish as temperature approaches 0 K. Between about −50 °C and 0 °C, the commonly quoted 2.09 value is a reasonable approximation, but anyone working outside that window should use a temperature-dependent model.
Pressure also matters, though less dramatically over ordinary ranges. Under extreme pressures, ice can adopt crystal structures quite different from the familiar hexagonal form (ice Ih) found in your freezer. These exotic forms, known as ice II through ice XIX and counting, have different densities and different thermal properties. The specific heat of these high-pressure phases is a subject of active research, mostly relevant to planetary science and materials physics rather than anything you would encounter in a kitchen.
Steam, Liquid, and Ice on the Same Chart
If you plot the specific heat of Hâ‚‚O across its three common phases, you get an interesting picture. Ice near 0 °C is at about 2.09, liquid water is at 4.18, and steam (water vapor at constant pressure near 100 °C) is at roughly 2.01 J/g·°C. The liquid phase is the outlier, with a specific heat nearly double that of either the solid or the gas. This is unusual. For most substances, the liquid phase has a specific heat between the solid and the gas, or all three are in a fairly narrow range. Water’s liquid phase is anomalously high because of its extensive, constantly reorganizing hydrogen-bond network, which provides an unusually large number of ways to absorb thermal energy.
The near-equality of ice and steam’s specific heats is partly coincidental, since the underlying physics is quite different. In steam, each water molecule moves independently and absorbs energy through translational and rotational motion. In ice, molecules vibrate collectively within a lattice. The two mechanisms happen to land on similar numerical values per gram, but they are physically unrelated processes. Liquid water’s dramatically higher value reflects its unique status as a substance where both molecular motion and an extensive bonding network contribute simultaneously.
When Snowpack Becomes a Thermal Battery
One of the more consequential real-world applications of ice’s specific heat is in snow hydrology and permafrost science. A seasonal snowpack is essentially a thick blanket of porous ice sitting on the ground for months. Its ability to absorb or release heat as air temperatures fluctuate above and below it depends directly on the specific heat of ice, along with the thermal conductivity and density of the snow layer.
Snow acts as an insulator partly because of the air trapped between ice crystals, but the ice crystals themselves serve as a thermal buffer. When air temperatures drop sharply, the snowpack cools, absorbing energy from the ground beneath it at a rate governed by ice’s specific heat. When temperatures rise, the snowpack warms before any melting starts, again buffered by the specific heat. This thermal inertia helps protect permafrost from rapid temperature swings and moderates soil temperatures for plant roots and burrowing animals. In climate models, the thermal properties of snow and ice are essential inputs, and the specific heat of ice is one of the most basic of those inputs, feeding into predictions of permafrost stability, spring melt timing, and regional water availability.
For anyone building an energy balance model of a frozen surface, the specific heat of ice is only one piece of a larger puzzle that includes thermal conductivity, albedo, latent heat, and the density of the snow or ice layer. But it is the piece that governs how much the temperature of the frozen material itself changes in response to a given energy input, and getting it right at the relevant temperature range, rather than defaulting to the 0 °C value, can measurably improve model accuracy in cold environments.