Jump to content Jump to footer

Freezing Processes

Sea ice isn’t “just” frozen water. The salts contained in seawater change its characteristics – and especially its freezing behaviour – considerably.

In this section, we’ll first take a look at the fundamental physical properties of water, which provide a basis for understanding sea-ice formation. In the next step, we’ll shed some light on the salinity of seawater. What is salt? Which salts are there in seawater? And how high is the water’s salinity in various ocean regions?

It’s also important to know how salinity affects the freezing point of seawater, why from a certain salinity it loses the unique quality of negative thermal expansion, and how this affects the freezing process in the ocean.

When seawater freezes, the water molecules displace the salt from the forming crystal structure, pushing it into the surrounding water. This gradually forms a complex system of cavities and channels in which high-saline brine collects. When the temperature drops in sea ice, part of the seawater freezes in the brine pores, making them smaller in the process, and the salinity of the liquid brine rises. How sea-ice composition depends on temperature can be clearly described using a phase diagram.

The unique crystalline structure of ice, clearly visible in the diagram, is e.g. responsible for the fact that, when pressure is applied from a certain direction, ice can be deformed, while it remains rigid in response to pressure applied from other directions. In addition, the crystalline structure influences how sea ice grows.

In the closing section, we will explain how ice crystals that form in the ocean rise to the surface, where they produce what is known as frazil ice. We will show (under Sea-ice Growth) how, depending on the weather, frazil ice can then transform into sheets of solid ice (nilas) or small floes (pancake ice), which – when they collide and stack due to drift processes – can form pressure ridges towering up to 50 metres tall.

Sea-ice Freezing Processes

Water (H2O) is the only substance on Earth that can be found in the three classical states solid, liquid and gaseous under our planet’s climatic conditions. When water changes states (phase transition), energy is transferred. Depending on the type of transition, energy is either absorbed and bound (melting, vaporisation) or released (condensation, freezing). The states and phase transitions of water can be represented in a pressure and temperature diagram. From the diagram (see the figure), we can see e.g. that at normal pressure (1013 hPa) the freezing point of water (= melting point of ice) lies at 0 °C and its boiling point lies at 100 °C. As pressure rises, the melting point decreases and boiling point increases. As pressure drops, the melting point increases, while the boiling point decreases. Accordingly, water will boil far below 100°C on a mountaintop. At the extremely low pressure of 6.1 hPa, the melting and vaporisation curves coincide at what is called the triple point. Here, water can be simultaneously found in all three states at a temperature of 0.01 °C. The table provides further information on the properties of water and their relevance.

The Open University (1995): Seawater: Its Composition, Properties and Behaviour, p. 6

Dissolved salts change the characteristics of water. The more salt seawater contains, the higher its density and the lower its freezing point. Accordingly, the salinity of seawater is of great importance with regard to sea-ice formation.

The salts dissolved in seawater are “dissociated”; that is, they manifest in the form of charged particles (ions). Approximately 99.9 percent of the salts found in seawater are formed by a handful of negatively charged molecules (anions) and positively charged particles (cations). The most important anions are chloride (Cl-), sulphate (SO4-2), bicarbonate (HCO3-), bromide (Br-) and boric acid dissociated in water (H2BO3-). The most important cations are sodium (Na+), magnesium (Mg2+), calcium (Ca2+), potassium (K+) and strontium (Sr2+).

In keeping with the principle of constant proportions, these ions can always be found in the same relative proportions, regardless of the seawater’s overall salinity. Consequently, the salinity of an entire ocean can be approximated using just one figure – the number of chloride ions.

As a rule, salinity is presented as the absolute salinity in percent (%) or parts per thousand (‰). In other words, water with a salinity of 1‰ contains 1 gram of salt per kilogram (i.e., per litre).

Salinity varies from ocean to ocean and is also influenced by ocean currents. The mean salinity of all oceans is 34.7 parts per thousand, with a relatively small distribution around the mean. 50 percent of all oceans have a salinity of between 34.6 and 34.8 parts per thousand. The average range of ocean salinity varies from 32 to 38 parts per thousand (Krauß, 2011). If the shelf seas, which can have comparatively high or low salinities, are included, the values vary from 28 to 40 parts per thousand.

The comparatively large deviations in the coastal marginal seas can be attributed to the effects of natural processes on salinity. These include evaporation, precipitation and freshwater input from the continents:

  • In the marginal seas of the temperate latitudes in the Northern and Southern Hemisphere alike, the salinity is lower because they receive comparatively large amounts of freshwater from rivers and precipitation.
  • Conversely, in the marginal seas of the subtropical latitudes, evaporation outweighs the freshwater input from rivers and precipitation, and in some cases the salinity is higher than in the open ocean.

In the Baltic Sea, the salinity varies considerably. Off the coast of Schleswig-Holstein, it is 15 to 18 parts per thousand. Far to the east, between Sweden and Finland, various rivers contribute freshwater to the Baltic. Here the salinity is only 3 to 5 parts per thousand. | In the Dead Sea, the salinity can reach 330 parts per thousand. The salt affects the water’s density to such an extent that swimmers can virtually float on its surface. | Arctic Ocean: 32 parts per thousand | Persian Gulf: up to 40 parts per thousand | Mediterranean: up to 39 parts per thousand | Red Sea: up to 41 parts per thousand | Atlantic Ocean: between 34.5 and 37 parts per thousand.

Krauß J, (2011):  Grundzüge der Maritime Meteorologie und Ozeanographie, Salzwasserverlag, Paderborn, p. 172

The relation between freezing point and maximum density at varying salinities is of great importance for the formation of sea ice.


Unlike freshwater, saltwater only freezes at temperatures below zero degrees Celsius. This is due to certain interactions between the salts and the water molecules. The saltier the water is, the lower the freezing point. This freezing-point decline is a linear progression (light blue line in the figure above). Similarly, as salinity increases, the temperature at which water reaches maximum density (dark blue line) changes.

However, as salinity increases, the maximum-density temperature sinks much more rapidly than does the freezing point. Accordingly, the two lines coincide at a temperature of ca. -1.3 degrees Celsius and a salinity of 24.7‰. At even higher salinities, the freezing point and the maximum-density temperature are always identical. Below the freezing point, no higher density can be reached because a certain percentage of the salts that seawater contains is lost during the freezing process, further reducing the density of the sea ice (Tardent, 2005).

As a result, seawater with a salinity of 24.7‰ or higher loses the unique quality of negative thermal expansion. Whereas nearly all fluids contract during freezing, increasing their density in the process, water expands when frozen. Consequently, pure water reaches its maximum density at 4 degrees Celsius, not at zero.

For example, when the surface of a freshwater lake cools in winter, the heaviest (4-degree-Celsius) water always sinks to the bottom, where it remains. The even colder water, with a temperature near zero, is at the surface. As a result, the lake freezes over up top, but remains liquid below.

In high-saline seawater, the process is a bit different. As it is not subject to negative thermal expansion, the coldest seawater is also the heaviest. When it cools on the surface, it sinks to the bottom of the upper mixed layer (ca. 50 to 100 metres thick), where it displaces warmer water, pushing the warm water upwards. In turn, it itself is displaced by even colder water from the surface and pushed upwards. This process ensures a constant mixing of the topmost layer, which continues until the entire layer has reached the freezing point.

Tardent P. (2005): Meeresbiologie, 3. unveränderte Auflage, Thieme-Verlag, pp. 176- 177

When seawater freezes, the salt is not encased in the ice crystals and stays in the remaining water. At high freezing rates, ice crystals grow on the surface so quickly that the salt in the seawater doesn’t have a chance to flow downward. The ice crystals then capture highly concentrated brine in the brine pores. If the temperature within the sea ice continues to drop, part of the seawater freezes in the brine pores, making them even smaller, while the salinity of the liquid brine rises.

How the brine pores – and with them, the structure of the sea ice as a whole – change in response to further cooling can perhaps best be described using a phase diagram. The diagram shows which states the briny solution in the pores is in at various temperatures and salinities (liquid solution, liquid solution + salt crystals, liquid solution + ice crystals, salt crystals + ice crystals).

The figure shows a simplified phase diagram for water (H2O) and salt dissolved within it (NaCl).

If the solution has a salinity of zero percent, it is pure water and freezes at 0 degrees Celsius. If salt is mixed with the pure water at a temperature of 10 degrees Celsius, it will continue to dissolve until a salinity of ca. 27 percent is reached (Point 1 in the figure). The solution is then saturated, and the remaining salt remains in crystal form.

How does this 27-percent saline solution behave when cooled? Solid salt gradually separates from the solution, because the saturation point shifts as the temperature drops (in the diagram: red line on the left). At -21 degrees Celsius (Point 2 in the figure), the solution freezes. When it freezes, the two solids separate, leaving mixed ice and salt crystals.

What happens when a 10-percent saline solution (Point 3 in the figure) at 10 degrees is cooled? Initially, the solution remains liquid and homogeneous. At a temperature of -7°C (Point 4 in the figure) the first ice crystals form, while the salt remains in the solution and becomes more concentrated. As a result, the salinity of the remaining solution continues to climb until a temperature of -21 degrees Celsius and salinity of 23.3% are reached (Point 5 in the figure). Then the entire system freezes, producing separate ice and salt crystals.

In the middle section of the diagram, there is only one phase (liquid homogeneous salt solution), while there are two each in the other three sections: on the right, a liquid salt solution with salt crystals; on the left, a liquid salt solution with ice crystals; and below, solid ice and salt crystals.

Consequently, when seawater begins to freeze, the percentage of water in the solution continually declines and the freezing point drops with it. This process continues until the solution is saturated with salt. The lowest temperature for a liquid saline solution is –21 degrees Celsius. At this temperature, salt from the solution (as NaCl * 2 H2O) begins to crystallise, along with the ice. The frozen solution is then a mixture of separate NaCl * 2 H2O crystals and ice crystals and no longer a homogeneous mixture of salt and water. This form of the phase diagram of saltwater describes what is known as a eutectic state. The phase diagram is characterised by one liquid phase and two different solid phases for the solid-to-liquid transition.

For sea ice, which can contain a diverse range of salts, the phase diagram is of course much more complex. At a temperature of -2.2 degrees Celsius, calcium carbonate separates. As temperatures continue to drop, other salts follow suit. The separation temperatures for specific salts in the solution are: -8.2 degrees Celsius for sodium sulphate, -22.9 degrees Celsius for sodium chloride, -36 degrees Celsius for KCl, and -54 degrees Celsius for CaCl.

In and of itself, a phase diagram offers no direct information on the concrete spatial configuration of individual phases within the “sea ice” system, e.g., on the microstructure of sea ice. In the case of natural sea ice, this depends on different factors: the environment in which the ice grows, and the conditions at the ice/water boundary layer, as well as the in-situ temperature and the chemical composition of the ice layer in question. The latter aspect is essential for a range of sea-ice properties, e.g. for the physical differences between ice, brine, salts and pockets of gas.

Clarke C. (2012): The science of ice cream, RSC Publishing, 2nd edition, pp..29-30

In its solid state – i.e., in the form of frozen ice – water has a crystalline structure. This structure has a significant influence on the properties of sea ice.

Depending on the pressure and temperature, ice can manifest in more than a dozen different variants, known as polymorphs. At normal atmospheric pressure and naturally occurring temperatures, water (H2O) crystallises in a hexagonal (six-sided) structure. Other types of ice are only found at extremely high pressures and/or extremely low temperatures.

In solid (ice) form, every water molecule is surrounded by four neighbouring molecules. They are arranged so that the oxygen atoms of the four neighbouring molecules form the corners of a nearly ideal tetrahedron. In turn, the oxygen atoms are bundled in a series of parallel planes, referred to as basal planes. The overall structure, reminiscent of a beehive, has a honeycomb pattern and is composed of layers of slightly irregular hexagons.

This crystalline structure influences many of sea ice’s physical and mechanical properties. For example, the behaviour of ice depends e.g. on the direction from which force is applied to it (anisotropy). The crystalline structure is also responsible for the formation of various classes of texture (e.g. columnar or granular ice) in sea ice.

During ice formation, at first tiny crystals appear, and often determine the further direction of growth for the ice. When an ice crystal grows, it costs a new atom less energy to dock on an existing basal plane than to start a new plane. As a result, ice is characterised by preferred directions of growth, which shape its appearance in a number of ways – e.g. in the branches of a star-shaped snowflake or in the fact that, in existing sea-ice cover, new crystals below the ice tend to grow downward, forming centimetre-long columns.

The preferred directions of growth are also what make ice anisotropic. If force is applied to its stacked layers laterally, the individual layers can shift and slide, causing the ice to deform. In contrast, if force is applied from any other direction, ice is far more rigid.

The crystalline structure of ice can fundamentally be described by the Bernal-Fowler Rules (Wadhams, 2000).

Bernal-Fowler rules

Principally speaking, the following applies to frozen water:

  1. Every position in the structure is filled by a water molecule, which is in turn surrounded by four neighbouring molecules, forming a tetrahedron.
  2. Between every oxygen pair (O-O) there is exactly one water molecule (H), which connects the two oxygen atoms via a hydrogen bond (O-H-O).
  3. The water molecules remain intact upon freezing, i.e., every oxygen atom bonds with two nearby hydrogen atoms (the O-H distance is ca. 1 angstrom (Å)) and is connected to two additional hydrogen atoms via a hydrogen bond.
  4. Since the lengths of the hydrogen bonds can vary, there are a range of possible configurations. This is why ice has an irregular structure.

Wadhams P. (2000): Ice in the ocean, Gordon and Breach Science Publisher, pp. 38-39

In winter, the cold atmosphere intensively cools the surface water on the ocean. As a result, its density increases, causing the surface water to sink, while warmer water from below “replaces” it at the surface. In turn, this water is cooled by the atmosphere and sinks. This produces constant vertical mixing, which continues until the entire upper water layer reaches the freezing point.

The upper mixed layer is typically between 50 and 100 metres thick but can also grow to several hundred metres in response to strong winds. Below it lie what are known as thermoclines, in which the temperature and salinity can rapidly change (thermoclines, haloclines). The mixing that goes on at the surface ends when it encounters these thermoclines.

Once the upper mixed layer has cooled to the freezing point, small ice crystals form in the water column. Because salt separates from the crystals during the freezing process, their density is lower than that of the surrounding water. As a result, they slowly rise to the surface.

The crystals initially form tiny platelets measuring only a few millimetres across. These then grow tiny branches, giving the crystals a star-like appearance. At this point, the formation of new crystals is greatly accelerated by a process referred to as collision breeding. The star-shaped ice particles collide with others of their kind, losing branches in the process. In turn, these fragments become the cores of new crystals.

This gradually produces a thin “slush” of ice crystals known as frazil ice, which steadily solidifies in the subsequent cooling and freezing process (Weeks, 1994).

Wadhams P., (2000):  Ice in the ocean, Gordon and Breach Science Publishers, p. 48
Weeks W. F. (1989): Growth conditions and the structure and properties of sea ice, in: Physics of the ice-covered seas, Volume 1, lecture notes from a summer school in Savonlinna, Finland 6-17 June 1994, Helsinki University Press, Helsinki, p. 39

Once the upper mixed layer of seawater reaches the freezing point, delicate ice crystals and ice platelets (“frazil ice”) form. They rise to the surface, where they appear as an icy “slush”. If the conditions on the ocean are calm, they eventually form a solid, smooth sheet of ice referred to as nilas. Under turbulent conditions, the slush turns into countless tiny floes that, given their appearance, are called pancake ice. Both nilas and pancake ice can subsequently grow into a solid sheet of young ice. Depending on how long it survives, this can be(come) a sheet of first-year ice with a thickness of between 30 and 150 centimetres, or of multiyear ice measuring up to several metres thick.


In many regions of the Arctic and Antarctic, the ocean is not completely covered with ice. There are often patches of open water amid the ice, some of which can be quite large; then they are referred to as polynyas. As a rule, they are produced by wind, tides, or rising warm seawater, can in some cases survive for years, and have a major influence on ocean currents, the climate and marine organisms. Near the coast, polynyas are chiefly produced by offshore winds, which drive the ice toward the open ocean (coastal polynyas). Far from the coast, polynyas are mostly formed by warm water rising from the ocean’s depths.

When extremely cold atmospheric conditions prevail for prolonged periods, sea ice can reach thicknesses of up to three metres. Ice floes and ice broken up by tides and wave action are constantly in motion due to winds and ocean currents (sea-ice drift). When two drift routes converge or come to a bottleneck (convergent drift), the floes can collide and stack up (rafting). This can produce what are known as pressure ridges, which can reach thicknesses of up to 50 metres.

Lubin D. & Massom R. (2006): Polar Remote Sensing, Volume 1, Atmosphere and Oceans, Springer Verlag, p. 318

D.N. Thomas D. & G. Dieckmann (2013): Sea ice: An introduction to its physics, chemistry, biology and geology, Blackwell Publishing, p. 23