What is 'unfreezable water'? Is it the same as bound water or water of hydration?

Equilibrium effects

The freezing point of water may be lowered by the osmotic effects of dissolved solutes or by the hydration effects of macromolecules or biological ultrastructures such as membranes.

The osmotic depression of freezing point has been known for a long time. For each mole of dissolved molecules or ions one adds to a kilogram of water, one lowers the equilibrium freezing point by (approximately) two degrees.

The depression of freezing point by hydration has only been studied quantitatively more recently. Water within about about one nanometre (0.000001 mm) of a hydrophilic surface (one that attracts water) freezes at a lower temperature than water unaffected by surfaces. The closer to the surface, the lower the freezing temperature. The 'last' few layers of water, the water molecules closest to the surface, are very difficult to freeze or to remove. This is partly due to the equilibrium hydration effect that we are currently discussing, and partly due to non-equilibrium effects discussed later. But the important points are that the equilibrium freezing point of water falls as the water content decreases and that this effect can be quantified.

Non-equilibrium effects

Supercooling is possible. Small volumes of water, or large volumes of very pure water, can be cooled considerably below 0C before they freeze. The water in biological systems is found in very small volumes, such as the interiors of cells and organelles, and so can supercool substantially. High viscosity hinders equilibration. In biological systems, solutes are present. When some of the water freezes, the solutes are concentrated in the remaining unfrozen water, producing highly concentrated solutions. These are viscous, and become more viscous at lower temperatures. This has a few consequences. One is that the molecules of water are slower to diffuse and to rotate, so that ice nuclei are less likely to form and supercooling is more likely. Another is that water cannot quickly diffuse from the region near a hydrophilic surface to a region where ice has already started to form. Thus the amount of unfrozen water is higher than one would expect at equilibrium.

High viscosity and low temperature can produce a glass. A glass is a non-crystalline solid. Window glass is an example. Glasses are formed when the viscosity of a liquid becomes so high that it can resist shear stresses (resist changes in shape) for extremely long periods. When the water in cells forms a glass, diffusion, freezing and biochemistry are virtually stopped.

How hard is 'water of hydration' bound? The term 'bound water' is often misunderstood and often used in a misleading way. Unless vitrification has occurred or unless the aqueous medium has a very large viscosity, water of hydration can exchange rather rapidly with any other nearby water molecules. However, even moderately elevated viscosity in the aqueous phase of a low hydration system can make equilibration very slow. Just because an experimentalist has difficulty removing water from a sample doesn't mean that it is bound. For further discussion, download this paper.

Which solutes are best at preventing freezing?

The answer to this FAQ usually involves several specific details, so I'll only make a few general comments. These are dealt with in more detail by Wolfe and Bryant (2001) (see references below).
    First, the solute should be soluble and, to achieve high concentrations, you may need high solubility. Supersaturation often occurs at freezing temperatures, but it helps if you start with high concentration. For osmotic freezing point depression, the direct effect of small solutes are rather similar at low concentration. Provided that one counts dissociating solutes (MgCl2 in solution is three solute ions), the freezing point depression is approximately proportional to concentration. At high concentration, the osmotic effects of salts may be less than proportional to concentration. Conversely, the osmotic effect of many solutes such as sugars increases at high concentration by more than simple proportionality.

    If the solute crystallises, this limits the concentration you can achieve, and hence the viscosity. So, to achieve large nonequilibrium effects, you need solutes that do not crystallise easily. One way to do this is to look through the handbooks for solutes that do not individually crystallise until very high concentration. Easier and often much cheaper is to use a mixture of two or more solutes.

    High viscosity helps, for reasons discussed above. Some solutes have a greater effect on viscosity than others. Large solute molecules often have a greater effect on viscosity than small.

    If you wish to prevent freezing in biological samples, you may also be interested in whether or not the solute is toxic. Also, you may want it to permeate the membrane. (If it doesn't the cell will contract osmotically, with potentially damaging physical and biochemical effects.) Solutes that permeate the membrane are usually toxic in high concentration, so some compromises are usually necessary: one may use a cocktail of permeating solutes and, after thawing, wash them out as quickly as possible.

Which solutes are best at preserving ultrastructure?

This is another subtle FAQ and it is complicated by the fact that some people in the cryopreservtion industry have a pecuniary interest in certain solutes. Further, the cytoplasm is a rather special environment, and ultrastructural elements (membranes and macromolecules) are often stabilised by effects that, in a sense, balance each other. Nevertheless, some general observations may be made.
  • Much ultrastructure is stabilised by the surface tension of water: it holds membranes together and is important in the secondary, tertiary and quaternary configuration of many proteins. So solutes that reduce substantially the surface tension of water (such as strong detergents) tend to disrupt ultrastructure.
  • Ions (particularly polyvalent ions) shield the electrostatic interaction,and also to some extent the dipolar interaction. Increasing the concentration of dissolved ions can reduce electrostatic repulsions and allow precipitation. So it's easy to imagine ultrastructural configurations that are sensitive to ion concentration.
  • Membrane soluble molecules are potentially dangerous. The stability of a lipid bilayer depends on the geometry of its hydrophobic region, so hydrophobic molecules can disrupt the semipermeability and even the bilayer structure. This is a serious problem in cryobiology, because often one wishes to have the solute permeate the membrane, yet one doesn't wish to disrupt it.
  • Some solutes interact with membranes, particularly at extreme dehydration. This statement is obvious: when one achieves a water content of say 20% or less, nearly all solute molecules are very near to a water-membrane or water-macromolecule interface.
  • Some molecules appear to affect the hydration forces between membranes (and may therefore affect the hydration forces between macromolecules). This is potentially important, because the the stresses produced by hydration forces at low water content can disrupt ulstrastructure. Yoon et al (1998) studied the effects of dimethylsulphoxide (DMSO), sorbitol, sucrose and trehalose in this regard. The effects of DMSO and sorbitol were consistent with what one would expect from their osmotic properties alone. Sucrose and trehalose had a greater effect in reducing the stress due to hydration forces.


For all of the reasons listed above, biological samples, or other samples in which water occurs in finely divided volumes, can exhibit freezing over a wide range of temperatures. A substantial amount of water may never freeze under given experimental conditions. This water may be very difficult to remove, but still not 'bound'. So what is 'unfreezable water'? The practical answer is that it is water that did not freeze when the experimenter expected that it would. This answer is unsatisfactory because it is not objective. So we prefer not to use the term at all. Instead, it is better to recognize that water can freeze over a wide range of temperatures, and that, if the temperature is cold enough, non-equilibrium effects will slow or prevent any further freezing. The 'water of hydration' that remains unfrozen may be considerably more than just a single layer of water molecules covering all hydrophilic surfaces. For quantitative estimates, and for a more formal discussion, see "What is 'unfreezable water', how unfreezable is it and how much is there?".

photo of snowgums

Snow gums in SE Australia

Some relevant references

More about water

Cryobiology links Acknowledgements
Joe Wolfe, J.Wolfe@unsw.edu.au, phone 61- 2-9385 4954 (UT + 10, +11 Oct-Mar).
School of Physics, University of New South Wales, Sydney, Australia.

 Joe's scientific home page
 Joe's educational pages
 Joe's music page

Pic of the author